CN109655751A - A kind of method and system using Gaussian process regression estimates battery charging state - Google Patents
A kind of method and system using Gaussian process regression estimates battery charging state Download PDFInfo
- Publication number
- CN109655751A CN109655751A CN201910148355.8A CN201910148355A CN109655751A CN 109655751 A CN109655751 A CN 109655751A CN 201910148355 A CN201910148355 A CN 201910148355A CN 109655751 A CN109655751 A CN 109655751A
- Authority
- CN
- China
- Prior art keywords
- unit
- gaussian process
- acquisition
- process regression
- signal
- 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
Landscapes
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
- Secondary Cells (AREA)
Abstract
The invention discloses a kind of method and systems using Gaussian process regression estimates battery charging state, comprising the following steps: the parameter in A, sensor group acquisition battery charging process, including charging temperature, charging voltage and charging current；B, optimization processing is carried out after the parameter of transducing signal acquisition unit acquisition sensor group acquisition, and digital signal is converted analog signals by AD conversion module；C, controller receives and constructs training sample set according to temperature value, voltage value, current value after digital signal；D, Gaussian process regression model out is learnt based on training sample set；E, estimated value is finally exported, the evaluation method that the present invention uses, which can be realized, carries out high-precision estimation to the charge parameter of battery charging state, and error is small, can effectively improve follow-up data treatment effeciency.
Description
Technical field
The present invention relates to battery charging state estimating system technical fields, specially a kind of to use Gaussian process regression estimates
The method and system of battery charging state.
Background technique
Charged state is defined as the percentage of charge available remaining in battery.SoC provides when cope with battery charging
Instruction, which can enable battery management system to improve electricity by making battery from overdischarge or overcharge situation
The pond service life.Therefore, it accurately measures SoC for suitable battery management and is very important.
Rechargeable battery carrys out storage energy by reversible chemical reaction.Traditionally, rechargeable battery provides lower use
Cost, and the result is that supporting towards the green proposal for influencing environment compared with non-rechargeable battery.For example, including consumer
In extensive application in electronic product, Roof of the house solar photovoltaic system, electric vehicle, smart electric grid system etc., lithium ion (Li
Ion) rechargeable battery has been widely deployed as main energy storage member.Li ion battery is more than to have different chemical reactions
Other kinds of battery at least some major advantages be low self-discharge rate, high single battery voltage, high-energy-density, lightweight,
Long-life and low-maintenance.
However, Li ion battery and other kinds of battery are chemical energy storage sources, and the chemical energy can not directly make
With.The problem leads to the SoC for being difficult to estimate battery.Because battery model can not capture nonlinear kinetics based on physics and
Associated parameter uncertainty, so the accurate estimation of SoC is still extremely complex and is difficult to carry out.Though current evaluation method
Charged state can so be estimated, but estimation process is considerably complicated, and error is big.
Summary of the invention
The purpose of the present invention is to provide a kind of method and system using Gaussian process regression estimates battery charging state,
To solve the problems mentioned in the above background technology.
To achieve the above object, the invention provides the following technical scheme: a kind of filled using Gaussian process regression estimates battery
The method of electricity condition, comprising the following steps:
A, the parameter in sensor group acquisition battery charging process, including charging temperature, charging voltage and charging current；
B, optimization processing is carried out after the parameter of transducing signal acquisition unit acquisition sensor group acquisition, and will by AD conversion module
Analog signal is converted to digital signal；
C, controller receives and constructs training sample set D, D={ T according to temperature value, voltage value, current value after digital signali, Vi,
Ii}n i=1；
D, Gaussian process regression model out is learnt based on training sample set；
E, estimated value is finally exported.
Preferably, optimization method is as follows in the step B:
A, the transducing signal in the battery charging process of acquisition is converted into signal matrix；
B, the Fourier transform of traveling every trade is pressed to signal matrix；
C, based on obtained row Fourier transform results, transducing signal signal filtering factor is iterated to calculate, to obtained row Fourier
Leaf transformation result is modified；
D, judge whether current iteration number is equal to the length of transducing signal sequence；If so, to currently available correction result,
The transducing signal sequence for having filtered out noise is generated by traveling every trade inverse fourier transform.
Preferably, Gaussian process regression model learning method is as follows in the step D:
A, the input x and x ' of any two sample is defined kernel function k (x, x ') are as follows:
, wherein x T For the transposition of x, δ
For Kronecker delta function, i.e.,, wherein σ f, c and σ n are
4 different hyper parameters in kernel function, and σ f is that signal standards is poor, c is yardstick adjustment factor, and σ n is noise criteria
Difference；
B, signal standards difference the σ f, yardstick adjustment factor c and noise criteria difference σ in hyper parameter are initialized:
c=0.2*σf；
C, observed value vector y=[y1, y2 ..., yn] is obtained according to training sample set T , and root
Covariance matrix is calculated according to kernel function k (x, x '):
；
D, value vector y and covariance matrix K y defines the log likelihood of training sample set according to the observation:
；
E, according to initialization hyper parameter, the log likelihood logp (y | D) of training sample set is maximized using conjugate gradient method
Optimal hyper parameter is obtained, determines Gaussian process regression model with these optimal hyper parameters.
Preferably, a kind of system using Gaussian process regression estimates battery charging state, including sensor group, sensing letter
Number acquisition unit, AD conversion unit, controller, evaluation unit, training sample set generation unit, is learned transducing signal optimization unit
Gaussian process regression model unit and output unit are practised, the sensor group input terminal connects rechargeable battery, the sensor group
Including temperature sensor, voltage sensor and current sensor, the temperature sensor acquires temperature when rechargeable battery work,
Voltage when the voltage sensor acquisition rechargeable battery work, the electricity when current sensor acquisition rechargeable battery works
Stream, the sensor group output end connect transducing signal by transducing signal acquisition unit and optimize unit, and the transducing signal is excellent
To change unit and controller is connected by AD conversion unit, the controller is separately connected evaluation unit, training sample set generation unit,
The evaluation unit, training sample set generation unit are all connected with study Gaussian process regression model unit, the study Gauss mistake
Journey regression model unit connects output unit.
Compared with prior art, the beneficial effects of the present invention are:
(1) evaluation method that the present invention uses, which can be realized, carries out high-precision estimation to the charge parameter of battery charging state, accidentally
Difference is small, can effectively improve follow-up data treatment effeciency.
(2) signal optimizing method that the present invention uses can effectively, quickly filter out the impulsive noise in transducing signal, into
One step improves subsequent estimation accuracy.
(3) the Gaussian process regression model learning method that the present invention uses can further increase estimation precision, reduce and miss
Difference.
Detailed description of the invention
Fig. 1 is flow chart of the present invention；
Fig. 2 is transducing signal optimization method flow chart of the present invention；
Fig. 3 is present system functional block diagram.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Fig. 1-3 is please referred to, the present invention provides a kind of technical solution: a kind of using Gaussian process regression estimates battery charging shape
The method of state, comprising the following steps:
A, the parameter in sensor group acquisition battery charging process, including charging temperature, charging voltage and charging current；
B, optimization processing is carried out after the parameter of transducing signal acquisition unit acquisition sensor group acquisition, and will by AD conversion module
Analog signal is converted to digital signal；
C, controller receives and constructs training sample set D, D={ T according to temperature value, voltage value, current value after digital signali, Vi,
Ii}n i=1；
D, Gaussian process regression model out is learnt based on training sample set；
E, estimated value is finally exported.
In the present invention, optimization method is as follows in step B:
A, the transducing signal in the battery charging process of acquisition is converted into signal matrix；
B, the Fourier transform of traveling every trade is pressed to signal matrix；
C, based on obtained row Fourier transform results, transducing signal signal filtering factor is iterated to calculate, to obtained row Fourier
Leaf transformation result is modified；
D, judge whether current iteration number is equal to the length of transducing signal sequence；If so, to currently available correction result,
The transducing signal sequence for having filtered out noise is generated by traveling every trade inverse fourier transform.
The signal optimizing method that the present invention uses can effectively, quickly filter out the impulsive noise in transducing signal, into one
Step improves subsequent estimation accuracy.
In the present invention, Gaussian process regression model learning method is as follows in step D:
A, the input x and x ' of any two sample is defined kernel function k (x, x ') are as follows:
, wherein x T For the transposition of x, δ
For Kronecker delta function, i.e.,, wherein σ f, c and σ n are
4 different hyper parameters in kernel function, and σ f is that signal standards is poor, c is yardstick adjustment factor, and σ n is noise criteria
Difference；
B, signal standards difference the σ f, yardstick adjustment factor c and noise criteria difference σ in hyper parameter are initialized:
c=0.2*σf；
C, observed value vector y=[y1, y2 ..., yn] is obtained according to training sample set T , and root
Covariance matrix is calculated according to kernel function k (x, x '):
；
D, value vector y and covariance matrix K y defines the log likelihood of training sample set according to the observation:
；
E, according to initialization hyper parameter, the log likelihood logp (y | D) of training sample set is maximized using conjugate gradient method
Optimal hyper parameter is obtained, determines Gaussian process regression model with these optimal hyper parameters.
The Gaussian process regression model learning method that the present invention uses can further increase estimation precision, reduce error.
In the present invention, a kind of system using Gaussian process regression estimates battery charging state, including sensor group, sensing
Signal acquisition unit 1, transducing signal optimization unit 2, AD conversion unit 3, controller 4, evaluation unit 5, training sample set generate
Unit 6, study Gaussian process regression model unit 7 and output unit 8, the sensor group input terminal connect rechargeable battery 9, institute
Stating sensor group includes temperature sensor 10, voltage sensor 11 and current sensor 12, and the acquisition of temperature sensor 10 is filled
Temperature when battery works, the voltage sensor 11 acquire voltage when rechargeable battery work, the current sensor 12
Electric current when rechargeable battery work is acquired, the sensor group output end connects transducing signal by transducing signal acquisition unit 1
Optimize unit 2, the transducing signal optimization unit 2 connects controller 4 by AD conversion unit 3, and the controller 4 is separately connected
Evaluation unit 5, training sample set generation unit 6, the evaluation unit 5, training sample set generation unit 6 are all connected with study Gauss
Process regression model unit 7, the study Gaussian process regression model unit 7 connect output unit 8.Sensor group acquires battery
Parameter in charging process, including charging temperature, charging voltage and charging current；Transducing signal acquisition unit acquires sensor group
Optimization processing is carried out after the parameter of acquisition, and digital signal is converted analog signals by AD conversion module；Controller receives
Training sample set is constructed according to temperature value, voltage value, current value after to digital signal；Learn Gauss mistake out based on training sample set
Journey regression model；Finally export estimated value.
In conclusion the charge parameter progress that the evaluation method that the present invention uses can be realized to battery charging state is high-precision
Degree estimation, error is small, can effectively improve follow-up data treatment effeciency.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of variations, modification, replacement can be carried out to these embodiments without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (4)
1. a kind of method using Gaussian process regression estimates battery charging state, it is characterised in that: the following steps are included:
A, the parameter in sensor group acquisition battery charging process, including charging temperature, charging voltage and charging current；
B, optimization processing is carried out after the parameter of transducing signal acquisition unit acquisition sensor group acquisition, and will by AD conversion module
Analog signal is converted to digital signal；
C, controller receives and constructs training sample set D, D={ T according to temperature value, voltage value, current value after digital signali, Vi,
Ii}n i=1；
D, Gaussian process regression model out is learnt based on training sample set；
E, estimated value is finally exported.
2. a kind of method using Gaussian process regression estimates battery charging state according to claim 1, feature exist
In: optimization method is as follows in the step B:
A, the transducing signal in the battery charging process of acquisition is converted into signal matrix；
B, the Fourier transform of traveling every trade is pressed to signal matrix；
C, based on obtained row Fourier transform results, transducing signal signal filtering factor is iterated to calculate, to obtained row Fourier
Leaf transformation result is modified；
D, judge whether current iteration number is equal to the length of transducing signal sequence；If so, to currently available correction result,
The transducing signal sequence for having filtered out noise is generated by traveling every trade inverse fourier transform.
3. a kind of method using Gaussian process regression estimates battery charging state according to claim 1, feature exist
In: Gaussian process regression model learning method is as follows in the step D:
A, the input x and x ' of any two sample is defined kernel function k (x, x ') are as follows:
, wherein x T For the transposition of x, δ
For Kronecker delta function, i.e.,, wherein σ f, c and σ n are
4 different hyper parameters in kernel function, and σ f is that signal standards is poor, c is yardstick adjustment factor, and σ n is noise criteria
Difference；
B, signal standards difference the σ f, yardstick adjustment factor c and noise criteria difference σ in hyper parameter are initialized:
c=0.2*σf；
C, observed value vector y=[y1, y2 ..., yn] is obtained according to training sample set T , and root
Covariance matrix is calculated according to kernel function k (x, x '):
；
D, value vector y and covariance matrix K y defines the log likelihood of training sample set according to the observation:
；
E, according to initialization hyper parameter, the log likelihood logp (y | D) of training sample set is maximized using conjugate gradient method
Optimal hyper parameter is obtained, determines Gaussian process regression model with these optimal hyper parameters.
4. a kind of system using Gaussian process regression estimates battery charging state, it is characterised in that: including sensor group, sensing
Signal acquisition unit (1), transducing signal optimization unit (2), AD conversion unit (3), controller (4), evaluation unit (5), training
Sample set generation unit (6), study Gaussian process regression model unit (7) and output unit (8), the sensor group input terminal
It connects rechargeable battery (9), the sensor group includes temperature sensor (10), voltage sensor (11) and current sensor
(12), temperature when temperature sensor (10) the acquisition rechargeable battery work, voltage sensor (11) the acquisition charging electricity
Voltage when pond works, electric current when current sensor (12) the acquisition rechargeable battery works, the sensor group output end
Transducing signal optimization unit (2) is connected by transducing signal acquisition unit (1), transducing signal optimization unit (2) passes through AD
Converting unit (3) connects controller (4), and the controller (4) is separately connected evaluation unit (5), training sample set generation unit
(6), the evaluation unit (5), training sample set generation unit (6) are all connected with study Gaussian process regression model unit (7), institute
State study Gaussian process regression model unit (7) connection output unit (8).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910148355.8A CN109655751A (en) | 2019-02-28 | 2019-02-28 | A kind of method and system using Gaussian process regression estimates battery charging state |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910148355.8A CN109655751A (en) | 2019-02-28 | 2019-02-28 | A kind of method and system using Gaussian process regression estimates battery charging state |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109655751A true CN109655751A (en) | 2019-04-19 |
Family
ID=66123863
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910148355.8A Pending CN109655751A (en) | 2019-02-28 | 2019-02-28 | A kind of method and system using Gaussian process regression estimates battery charging state |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109655751A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110515001A (en) * | 2019-09-07 | 2019-11-29 | 创新奇智(广州)科技有限公司 | A kind of two stages battery performance prediction technique based on charge and discharge |
| CN110703113A (en) * | 2019-10-14 | 2020-01-17 | 重庆大学 | Power battery pack SOC estimation method based on Gaussian process regression |
| CN111460382A (en) * | 2020-03-30 | 2020-07-28 | 上海交通大学 | Fuel vehicle harmful gas emission prediction method and system based on Gaussian process regression |
| CN111460380A (en) * | 2020-03-30 | 2020-07-28 | 上海交通大学 | Multi-working-condition driving range prediction method and system based on Gaussian process regression |
| CN111460381A (en) * | 2020-03-30 | 2020-07-28 | 上海交通大学 | Multi-working-condition fuel vehicle oil consumption prediction method and system based on Gaussian process regression |
| CN114325398A (en) * | 2021-11-08 | 2022-04-12 | 淮阴工学院 | Fault detection method for proton exchange membrane fuel cell system |
| WO2023284167A1 (en) * | 2021-07-16 | 2023-01-19 | 苏州浪潮智能科技有限公司 | Battery charging method and apparatus, and device and medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550989A (en) * | 2015-12-09 | 2016-05-04 | 西安电子科技大学 | Image super-resolution method based on nonlocal Gaussian process regression |
| CN106339755A (en) * | 2016-08-29 | 2017-01-18 | 深圳市计量质量检测研究院 | Lithium battery SOH (State of Health) prediction method based on neural network and periodic kernel functions GPR |
| CN107422269A (en) * | 2017-06-16 | 2017-12-01 | 上海交通大学 | A kind of online SOC measuring methods of lithium battery |
| US20170350944A1 (en) * | 2016-06-06 | 2017-12-07 | Mitsubishi Electric Research Laboratories, Inc. | Methods and Systems for Data-Driven Battery State of Charge (SoC) Estimation |
| CN108918932A (en) * | 2018-09-11 | 2018-11-30 | 广东石油化工学院 | Power signal adaptive filter method in load decomposition |
-
2019
- 2019-02-28 CN CN201910148355.8A patent/CN109655751A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550989A (en) * | 2015-12-09 | 2016-05-04 | 西安电子科技大学 | Image super-resolution method based on nonlocal Gaussian process regression |
| US20170350944A1 (en) * | 2016-06-06 | 2017-12-07 | Mitsubishi Electric Research Laboratories, Inc. | Methods and Systems for Data-Driven Battery State of Charge (SoC) Estimation |
| CN109196366A (en) * | 2016-06-06 | 2019-01-11 | 三菱电机株式会社 | Use the method and system of Gaussian process regression estimates battery charging state |
| CN106339755A (en) * | 2016-08-29 | 2017-01-18 | 深圳市计量质量检测研究院 | Lithium battery SOH (State of Health) prediction method based on neural network and periodic kernel functions GPR |
| CN107422269A (en) * | 2017-06-16 | 2017-12-01 | 上海交通大学 | A kind of online SOC measuring methods of lithium battery |
| CN108918932A (en) * | 2018-09-11 | 2018-11-30 | 广东石油化工学院 | Power signal adaptive filter method in load decomposition |
Non-Patent Citations (1)
| Title |
|---|
| GOZDE OZCAN ET AL.: ""Online State of Charge Estimation for Lithiu-Ion Batteries Using Gaussian Process Regression"", 《IECON 2016 - 42ND ANNUAL CONFERENCE OF THE IEEE INDUSTRIAL ELECTRONICS SOCIETY》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110515001A (en) * | 2019-09-07 | 2019-11-29 | 创新奇智(广州)科技有限公司 | A kind of two stages battery performance prediction technique based on charge and discharge |
| CN110703113A (en) * | 2019-10-14 | 2020-01-17 | 重庆大学 | Power battery pack SOC estimation method based on Gaussian process regression |
| CN111460382A (en) * | 2020-03-30 | 2020-07-28 | 上海交通大学 | Fuel vehicle harmful gas emission prediction method and system based on Gaussian process regression |
| CN111460380A (en) * | 2020-03-30 | 2020-07-28 | 上海交通大学 | Multi-working-condition driving range prediction method and system based on Gaussian process regression |
| CN111460381A (en) * | 2020-03-30 | 2020-07-28 | 上海交通大学 | Multi-working-condition fuel vehicle oil consumption prediction method and system based on Gaussian process regression |
| CN111460381B (en) * | 2020-03-30 | 2022-03-18 | 上海交通大学 | Multi-working-condition fuel vehicle oil consumption prediction method and system based on Gaussian process regression |
| CN111460382B (en) * | 2020-03-30 | 2022-03-18 | 上海交通大学 | Fuel vehicle harmful gas emission prediction method and system based on Gaussian process regression |
| WO2023284167A1 (en) * | 2021-07-16 | 2023-01-19 | 苏州浪潮智能科技有限公司 | Battery charging method and apparatus, and device and medium |
| US11823025B1 (en) | 2021-07-16 | 2023-11-21 | Inspur Suzhou Intelligent Technology Co., Ltd. | Battery charging method and apparatus, and device and medium |
| CN114325398A (en) * | 2021-11-08 | 2022-04-12 | 淮阴工学院 | Fault detection method for proton exchange membrane fuel cell system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109655751A (en) | A kind of method and system using Gaussian process regression estimates battery charging state | |
| CN103616647B (en) | A kind of estimation method of battery dump energy for cell management system of electric automobile | |
| CN109031153A (en) | A kind of health status On-line Estimation method of lithium ion battery | |
| CN107422266B (en) | Fault diagnosis method and device for high-capacity battery energy storage system | |
| CN112540298B (en) | Lithium ion battery SOC estimation method based on deep-migration learning | |
| CN106026260A (en) | SOC estimation method for series-wound battery pack having equalization circuit | |
| CN111812515A (en) | XGboost model-based lithium ion battery state of charge estimation | |
| CN110806541B (en) | AD-BAS-based lithium battery model parameter identification method | |
| CN104714188A (en) | Method and system for estimating measured noise variance matrix matched battery state of charge (SOC) | |
| CN109633453A (en) | Battery parameter on-line identification method, apparatus and computer readable storage medium | |
| CN111060824A (en) | Charge state estimation method based on model and algorithm step-by-step fusion | |
| CN113702843A (en) | Lithium battery parameter identification and SOC estimation method based on suburb optimization algorithm | |
| WO2024000756A1 (en) | Electrochemical model parameter identification method based on deep learning and heuristic algorithm | |
| CN114779103A (en) | Lithium ion battery SOC estimation method based on time-lag convolutional neural network | |
| CN110749832A (en) | Method for quickly estimating actual capacity of retired lithium ion battery of electric vehicle | |
| Takyi-Aninakwa et al. | A hybrid probabilistic correction model for the state of charge estimation of lithium-ion batteries considering dynamic currents and temperatures | |
| CN112782594B (en) | Method for estimating SOC (state of charge) of lithium battery by data-driven algorithm considering internal resistance | |
| CN108693483A (en) | A kind of computational methods and system of battery pack state-of-charge | |
| CN105974320A (en) | Liquid or semi-liquid metal-cell state-of-charge estimation method | |
| CN106169910B (en) | Photovoltaic cell parameter identification method based on group hunting algorithm | |
| CN110361653A (en) | A kind of SOC estimation method and system based on hybrid accumulator | |
| CN115081332B (en) | Working condition sensitivity analysis and data processing method and device for parameter identification | |
| CN111060822A (en) | State of charge estimation method based on model switching and fusion | |
| CN113406503A (en) | Lithium battery SOH online estimation method based on deep neural network | |
| CN116707331A (en) | Inverter output voltage high-precision adjusting method and system based on model prediction |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190419 |
|
| WD01 | Invention patent application deemed withdrawn after publication |