CN112865221A - Automobile lithium battery dynamic charging protection system based on experience function and Internet of things - Google Patents
Automobile lithium battery dynamic charging protection system based on experience function and Internet of things Download PDFInfo
- Publication number
- CN112865221A CN112865221A CN202110014866.8A CN202110014866A CN112865221A CN 112865221 A CN112865221 A CN 112865221A CN 202110014866 A CN202110014866 A CN 202110014866A CN 112865221 A CN112865221 A CN 112865221A
- Authority
- CN
- China
- Prior art keywords
- charging
- battery
- threshold voltage
- information
- network model
- 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
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 title claims abstract description 46
- 229910052744 lithium Inorganic materials 0.000 title claims abstract description 46
- 238000003062 neural network model Methods 0.000 claims abstract description 57
- 238000005516 engineering process Methods 0.000 claims abstract description 7
- 238000004146 energy storage Methods 0.000 claims description 44
- 238000000034 method Methods 0.000 claims description 33
- 230000008569 process Effects 0.000 claims description 15
- 230000001133 acceleration Effects 0.000 claims description 5
- 230000006855 networking Effects 0.000 claims 1
- 230000003044 adaptive effect Effects 0.000 abstract description 4
- 230000002035 prolonged effect Effects 0.000 abstract description 4
- 230000004927 fusion Effects 0.000 abstract 1
- 230000006870 function Effects 0.000 description 49
- 230000000875 corresponding effect Effects 0.000 description 22
- 230000000694 effects Effects 0.000 description 7
- 230000001364 causal effect Effects 0.000 description 6
- 230000008859 change Effects 0.000 description 6
- 238000010845 search algorithm Methods 0.000 description 5
- 238000007599 discharging Methods 0.000 description 3
- 230000001960 triggered effect Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000009191 jumping Effects 0.000 description 2
- 230000001502 supplementing effect Effects 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 239000002253 acid Substances 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013398 bayesian method Methods 0.000 description 1
- 238000013477 bayesian statistics method Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000002922 simulated annealing Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J7/00—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries
- H02J7/0029—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries with safety or protection devices or circuits
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02H—EMERGENCY PROTECTIVE CIRCUIT ARRANGEMENTS
- H02H7/00—Emergency protective circuit arrangements specially adapted for specific types of electric machines or apparatus or for sectionalised protection of cable or line systems, and effecting automatic switching in the event of an undesired change from normal working conditions
- H02H7/18—Emergency protective circuit arrangements specially adapted for specific types of electric machines or apparatus or for sectionalised protection of cable or line systems, and effecting automatic switching in the event of an undesired change from normal working conditions for batteries; for accumulators
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J7/00—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J7/00—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries
- H02J7/0047—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries with monitoring or indicating devices or circuits
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J7/00—Circuit arrangements for charging or depolarising batteries or for supplying loads from batteries
- H02J7/34—Parallel operation in networks using both storage and other dc sources, e.g. providing buffering
- H02J7/342—The other DC source being a battery actively interacting with the first one, i.e. battery to battery charging
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Secondary Cells (AREA)
Abstract
According to the automobile lithium battery dynamic charging protection system based on the experience function and the Internet of things, the first charging threshold voltage and the second charging threshold voltage are dynamically generated by utilizing the neural network model and the experience function, the first charging threshold voltage and the second charging threshold voltage can be adjusted according to the loss condition of the batteries, meanwhile, the output result of the neural network model is corrected through the weighting factor, the output result can be dynamically attached to the actual condition of each battery, dynamic protection is more accurate and adaptive, the service life of the battery is greatly prolonged, in addition, the Internet of things technology is adopted, battery information is accurately transmitted in real time through fusion of various wired and wireless networks and the Internet, and charging information is more convenient and intelligent to obtain.
Description
Technical Field
The invention relates to the technical field of dynamic protection of battery charging, in particular to a dynamic charging protection system of an automobile lithium battery based on an empirical function and the Internet of things.
Background
At present, batteries (such as lithium batteries of mobile phones, lithium batteries of automobiles, lead-acid batteries, and the like) generally have a charging protection function and a quick charging function, but the current charging scheme generally adopts the quick charging function to charge the batteries to about 80%, and then adopts a slow charging mode to perform charging protection, so that the charging protection can be performed on the rechargeable batteries.
But on the one hand, the functions only carry out relatively rigid protection on the rechargeable battery, the loss of the rechargeable battery is not considered, the charging protection error is relatively large, and the battery cannot be effectively charged and protected, on the other hand, for the automobile lithium battery, the charging pile is required to be adopted for charging, the method increases the burden of the charging pile, particularly for the current popular charging pile, the charging pile has the electricity storage function, when the automobile lithium battery is charged, the charging voltage needs to be adjusted at any time, the charging power is adjusted at any time, the charging pile is undoubtedly damaged, the current charging protection is not complete fundamentally, the risk of the charging protection for the automobile lithium battery is only transferred to a charging pile manufacturer from a consumer end, and the problem of the charging protection is not solved.
Disclosure of Invention
To solve the problems in the prior art, an embodiment of a first aspect of the present invention provides a battery charging dynamic protection method, including:
acquiring charging information of a battery, wherein the charging information comprises residual recyclable time data, charging time data and charging duration information each time;
inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function; the neural network model is obtained by utilizing the charging information training of different batteries of the same type;
in the process that the battery is charged by a charging power supply, if the charging voltage of the battery is lower than the first charging threshold voltage, triggering to generate a first pulse current, and if the charging voltage of the battery reaches the second charging threshold voltage, triggering to generate a second pulse current; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the battery, and the second pulse current is used for triggering the energy storage device to be coupled with the charging power supply.
In a preferred embodiment, generating the second charging threshold voltage by the charging information and the set loss empirical function includes:
based on the ratio of the residual recyclable time data to the set total recyclable time data, an overshoot protection threshold and a fast-rush acceleration interval, generating an overshoot protection threshold and a fast-rush interval corresponding to the current residual recyclable time;
generating a threshold error correction value according to the charging time data, the charging time information of each time and a set loss empirical function;
and correcting the overshoot protection threshold value by using the threshold error correction value to obtain the second charging threshold voltage.
In a preferred embodiment, further comprising: and establishing the neural network model.
In a preferred embodiment, the neural network model is a bayesian network model.
In a preferred embodiment, further comprising:
acquiring scoring data corresponding to the total charging time from an expert database by combining with an expert model;
determining a corresponding weight factor according to the scoring data in combination with a preset corresponding relation table of the scoring data and the weight factor;
and correcting the output result of the Bayesian network model by using the weight factor.
An embodiment of a second aspect of the present invention provides a battery charging dynamic protection apparatus, including:
the charging information acquisition module is used for acquiring charging information of the battery, wherein the charging information comprises residual recyclable time data, charging time data and charging duration information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function;
the charging protection module triggers to generate a first pulse current if the charging voltage of the battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the battery reaches the second charging threshold voltage in the charging process of the battery through a charging power supply; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the battery, and the second pulse current is used for triggering the energy storage device to be coupled with the charging power supply.
In a preferred embodiment, further comprising:
and the neural network model generating module is used for establishing the neural network model.
In a preferred embodiment, further comprising:
the scoring data acquisition module is used for acquiring scoring data corresponding to the total charging time from the expert database in combination with the expert model;
the weighting factor generation module is used for determining a corresponding weighting factor according to the scoring data in combination with a preset corresponding relation table of the scoring data and the weighting factor;
and the result correction module corrects the output result of the Bayesian network model by using the weight factor.
An embodiment of a third aspect of the present invention provides a dynamic charging protection system for an automotive lithium battery, including: battery dynamic protection device and car charging pile, wherein battery charging dynamic protection device includes:
the charging information acquisition module is used for acquiring charging information of the automobile lithium battery from an automobile BMS system, wherein the charging information comprises residual recyclable time data, charging time data and charging time information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function;
the charging protection module triggers to generate a first pulse current if the charging voltage of the automobile lithium battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the automobile lithium battery reaches the second charging threshold voltage in the process that the automobile lithium battery is charged through the automobile charging pile; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the automobile lithium battery, and the second pulse current is used for triggering the energy storage device to be coupled with the automobile charging pile.
An embodiment of a fourth aspect of the present invention provides an automotive lithium battery dynamic charging protection system based on an empirical function and an internet of things, including: battery dynamic protection device and car charging pile, wherein battery charging dynamic protection device includes:
the charging information acquisition module is used for acquiring charging information of the automobile lithium battery from an automobile BMS system through the Internet of things technology, wherein the charging information comprises residual recyclable time data, charging time data and charging time length information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function;
the charging protection module triggers to generate a first pulse current if the charging voltage of the automobile lithium battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the automobile lithium battery reaches the second charging threshold voltage in the process that the automobile lithium battery is charged through the automobile charging pile; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the automobile lithium battery, and the second pulse current is used for triggering the energy storage device to be coupled with the automobile charging pile.
Advantageous effects
According to the automobile lithium battery dynamic charging protection system based on the empirical function and the Internet of things, the charging information is input into a preset neural network model, the neural network model outputs a first charging threshold voltage, a second charging threshold voltage is generated through the charging information and a set loss empirical function, further, when the charging voltage of a battery is in a lower level (lower than the first charging threshold voltage), the charging speed is improved by discharging and supplementing charging through an energy storage device, and when the charging voltage of the battery is in a higher level (higher than the second charging threshold voltage), the charging electric quantity of a charging power supply can be shared by charging through the energy storage device, so that the loss of the battery due to quick charging in the final end charging link is avoided, and the effect of dynamic balance charging is achieved; according to the technical scheme provided by the invention, the first charging threshold voltage and the second charging threshold voltage are dynamically generated by utilizing a neural network model and an empirical function, the first charging threshold voltage and the second charging threshold voltage can be adjusted according to the self-loss condition of the battery, and meanwhile, the output result of the neural network model is corrected through a weighting factor, so that the output result can be dynamically attached to the self-actual condition of each battery, the dynamic protection is more accurate and adaptive, the service life of the battery is greatly prolonged, in addition, the battery information is accurately transmitted in real time by adopting the technology of the internet of things and fusing various wired and wireless networks with the internet, and the charging information is more convenient and intelligent to obtain.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 is a schematic flow chart illustrating a dynamic protection method for battery charging according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a dynamic protection apparatus for battery charging according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a battery charging dynamic protection system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, a first embodiment of the present invention provides a battery charging dynamic protection method, including:
s101: acquiring charging information of a battery, wherein the charging information comprises residual recyclable time data, charging time data and charging duration information each time;
s102: inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function; the neural network model is obtained by utilizing the charging information training of different batteries of the same type;
s103: in the process that the battery is charged by a charging power supply, if the charging voltage of the battery is lower than the first charging threshold voltage, triggering to generate a first pulse current, and if the charging voltage of the battery reaches the second charging threshold voltage, triggering to generate a second pulse current; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the battery, and the second pulse current is used for triggering the energy storage device to be coupled with the charging power supply.
According to the technical scheme, the charging information is input into a preset neural network model, the neural network model outputs a first charging threshold voltage, a second charging threshold voltage is generated through the charging information and a set loss empirical function, further, when the charging voltage of the battery is in a lower level (the first charging threshold voltage), the charging speed is improved through discharging and supplementary charging of an energy storage device, when the charging voltage of the battery is in a higher level (the second charging threshold voltage), the charging electric quantity of a charging power supply can be shared by charging of the energy storage device, further, the loss of the battery due to quick charging of a final end charging link is avoided, the effect of dynamic balance charging is achieved, and meanwhile, the charging threshold voltage is dynamically generated based on the charging information, the empirical function and the neural network, the battery protection device can be adjusted according to the loss condition of the battery, and can be dynamically attached to the actual condition of each battery, so that dynamic protection is more accurate and adaptive, and the service life of the battery is greatly prolonged.
Specifically, in the implementation process of the present invention, when charging is started, the charging power supply charges the battery of the present invention with a constant current, and since in the initial stage of charging, rapid charging can be performed, at this time, since the charging voltage of the battery is lower than a second threshold voltage (for example, for a battery with a rated voltage of 48v, the second threshold voltage is 10v), a first pulse current is triggered, the first pulse current is used for triggering an energy storage device to be coupled to the battery, for example, the first pulse current triggers a switch to be turned off, the switch is coupled between the energy storage device and the battery, and thus the energy storage device and the battery can be disconnected and connected through on and off control of the switch.
When the energy storage device is coupled with the battery, because the energy storage device reaches saturation (namely the energy storage device is gradually in a charged state) under the second charging threshold voltage when the energy storage device is charged at the previous time, after the energy storage device is coupled with the battery, the energy storage device can serve as a supplementary power supply to feed the battery, and the charging speed of the battery is improved.
It can be understood that the present charging voltage of the battery in the present invention is related to the internal electric quantity of the battery and is positively correlated, when the battery is charged, the present charging voltage gradually increases from 0 to the maximum saturation voltage of the battery, and the present charging voltage may have one or more different terms in the industry, but the meanings thereof are consistent and are all parameters that can represent the present electric quantity, which is not described in the present invention.
When the current charging voltage of the battery reaches the second charging threshold voltage, the battery is already in a 'large-capacity' state, the battery can be damaged by quick charging at the moment, the prior art realizes the purpose by setting charging power (reducing the charging power), and then certain burden is caused to the charging source, especially for the automobile charging pile.
When the battery reaches the second charging threshold voltage, the second pulse current can be triggered, the energy storage device can be triggered to be coupled with the charging power supply by the second pulse current, and it can be understood that the voltage of the energy storage device is smaller than the voltage of the charging power supply, so that the energy storage device is in a charged state, the charging power supply charges the energy storage device, on one hand, the energy storage device shares the voltage of the charging power supply, so that the current charging power of the battery is reduced, on the other hand, the charging power supply is not required to perform power switching, so that the charging power supply is protected.
Furthermore, the charging information is input into a preset neural network model, the neural network model outputs a first charging threshold voltage, and a second charging threshold voltage is generated through the charging information and a set loss empirical function, namely the first charging threshold voltage and the second charging threshold voltage are variable and not fixed values, so that loss influence in the using process of the battery is avoided.
Specifically, in a preferred embodiment, generating the second charging threshold voltage by the charging information and the set loss empirical function includes:
based on the ratio of the residual recyclable time data to the set total recyclable time data, an overshoot protection threshold and a fast-rush acceleration interval, generating an overshoot protection threshold and a fast-rush interval corresponding to the current residual recyclable time;
generating a threshold error correction value according to the charging time data, the charging time information of each time and a set loss empirical function;
and correcting the overshoot protection threshold value by using the threshold error correction value to obtain the second charging threshold voltage.
In the embodiment, firstly, historical charging condition data of the battery is utilized, the consumption of the battery is considered, and the initially set overshoot protection threshold value and the fast charging acceleration interval are subjected to proportion conversion based on the ratio of the residual circulatable time data to the total circulatable time data to obtain the overshoot protection threshold value and the fast charging interval corresponding to the current residual circulatable time; and then combining the charging data with a set loss empirical function, wherein the different capacitance loss speeds can be caused by different charging habits, and further combining the loss empirical function to generate a threshold error correction value, wherein the correction value can correct the obtained process protection threshold, and finally, a second charging threshold voltage which is smaller in error and can be dynamically updated is obtained.
In a preferred embodiment, the set loss empirical function is a fitting function between the total charging time, the charging times data and the time length information of each charging and the threshold error correction value.
Specifically, the set loss empirical function is:
0.78846x total charging duration-0.73775 x number of charges x (median charging duration per charge).
In the embodiment, the correction value obtained by combining the charging time and the fitting of each charging information can be more accurately fitted and predicted to obtain the correction value in the current battery state, and the correction value is changed along with the change of the charging time and is also dynamically changed according to the fitting result.
In a preferred embodiment, further comprising: and establishing the neural network model.
Further, the neural network model is a bayesian network model.
As will be understood by those skilled in the art, the bayesian network model includes a topology of the bayesian network and a corresponding conditional probability table, and the topology of the bayesian network is used to represent a corresponding relationship between the charging information and the second charging threshold voltage.
Specifically, the bayesian network bn (bayesian network model), also called belief network, is composed of a Directed Acyclic Graph (DAG) and a Conditional Probability Table (CPT). In a bayesian network, two variables X and Y, if directly connected, indicate a direct dependency between them, and knowledge of X affects the confidence in Y and vice versa. In this sense, we mean that information can be passed between two directly connected nodes. On the other hand, if two variables X and Y are not directly connected, then information needs to pass between the two through the other variables. If all the information paths between X and Y are blocked, information cannot be passed between them. In this case, knowledge of one of the variables does not affect the confidence of the other variable, so X and Y are conditionally independent of each other. If the basic case is considered where two variables X and Y are indirectly connected through a third variable Z, the bayesian network can be decomposed into three basic structures, namely, forward, split and aggregate.
Among them, the advantages of the bayesian network are mainly reflected in:
(1) the Bayesian network describes the interrelation among the data by using a graph method, has clear semantics and is easy to understand. The graphical knowledge representation method facilitates the consistency and the integrity of the probability knowledge base, and the network module can be conveniently reconfigured according to the change of conditions.
(2) Bayesian networks are prone to handling incomplete data sets. All possible data inputs must be known for the traditional standard supervised learning algorithm, if some input is missing, the established model is biased, the method of the Bayesian network reflects a probability relation model among data in the whole database, and an accurate model can still be established if some data variable is missing.
(3) Bayesian networks allow learning causal relationships between variables. In the past data analysis, the causal relationship of a problem is that when the interference is large, the system can not make an accurate prediction. This causal relationship has been included in bayesian network models. The Bayesian method has causal and probabilistic semantics, and can be used for learning causal relationships in data and learning according to the causal relationships.
(4) The combination of Bayesian network and Bayesian statistics can make full use of domain knowledge and sample data information. The Bayesian network expresses the dependency relationship among variables by arcs, expresses the strength of the dependency relationship by a probability distribution table, organically combines the prior information with the sample knowledge, promotes the integration of the prior knowledge and the data, and is particularly effective when the sample data is sparse or the data is difficult to obtain.
Two main problems to be solved for establishing the bayesian network structure are respectively selection of a scoring function and selection of a searching method, which are specifically as follows:
(1) determining a scoring function:
and determining a scoring function corresponding to the Bayesian network according to the training sample set.
The commonly used scoring function is based on information theoretic criteria, which equates the learning problem to a data compression task, the learning objective being to find a model that describes the training data in the shortest code length, which in this case includes the byte length needed to describe the model itself and the byte length needed to describe the data using the model. For Bayesian network learning, the model is a Bayesian network, and each Bayesian network describes a probability distribution on training data, and a set of coding mechanisms can make the frequently-occurring samples shorter. Therefore, the bayesian network with the shortest overall coding Length (including the Description network and the coded data) should be selected, which is the Minimum Description Length (MDL) criterion.
Given training set D ═ x1,x2...,xmA bayesian network B ═ B<G,θ>The scoring function on D can be written as:
s(B|D)＝f(θ)|B|-LL(B|D) (1)
in the formula (1), | B | is the number of parameters of the bayesian network; f (theta) represents the number of bytes required to describe each parameter theta; therein
Is the log-likelihood of the bayesian network B. Obviously, the first term f (θ) | B | of equation (1) is the number of bytes required to compute the coded bayesian network, and the second term LL (B | D) is the probability distribution P corresponding to BBHow many bytes are needed to describe D. Thus, the learning task is transformed into an optimization taskThe objective is to find a Bayesian network B that minimizes the scoring function s (B | D).
If f (θ) is 1, that is, each parameter is described by 1 byte, the akaike information criterion scoring function AIC (B | D) is obtained as follows:
AIC(B|D)＝|B|-LL(B|D)
if it is
I.e. for each parameter
Byte description, the Bayesian Information rule BIC (Bayesian Information criteria) scoring function BIC (B | D) is obtained as follows:
obviously, if f (θ) is 0, i.e. the length of encoding the network is not calculated, the scoring function degenerates to negative log-likelihood and, correspondingly, the learning task degenerates to maximum likelihood estimation.
(2) And (3) searching algorithm:
with the scoring function determined, the learning problem of the bayesian network becomes a search problem. The search algorithm is to search for a bayesian network structure with the highest score under a certain scoring function. As the number of variables increases, the search space will increase at an exponential level with the number of nodes, finding the optimal model is the existence of a Non-Deterministic problem NP (Non-Deterministic polymeric Problems) that the Polynomial algorithm can solve. Heuristic search such as greedy search, simulated annealing, optimal first search and the like is commonly adopted at present.
The most common search method is to continuously change the directed edge in the network structure and judge the influence of each change on the score. If a directed edge exists between the two variables, the changing direction can be deleting the directed edge or reversing the directed edge; if there is no directed edge between two variables, the change method may be to add a directed edge in any direction, but when changing, a directed loop cannot be generated.
The simplest Search algorithm is a Greedy Search (Greedy Search). Let E denote the set of all candidate edges that may be added to the network structure, and Δ (E) denotes the variation of the scoring function after the edge E in E is added to the network structure. The search algorithm can be described as:
1) selecting an initial network structure;
2) selecting an edge E in the candidate edge set, so that delta (E) > delta (E '), wherein E' is any edge except for E in E, and delta (E) >0, stopping if an edge which meets the condition cannot be found, and turning to 3 if the edge does not meet the condition;
3) adding E to the network structure, deleting the edge from the candidate set E, and turning to 2);
in the algorithm, the initial network structure may be an empty network, a random network, or a priori network built using empirical knowledge. The greedy search strategy is a local search strategy and has the problem of trapping in local extrema and saddle points. One solution is to randomly change the structure of the network when a local extremum or saddle point is trapped, possibly jumping out of the saddle point or jumping from one local extremum region to another.
(3) Determining the topological structure of the Bayesian network based on a scoring function and a search algorithm:
the resulting bayesian network topology (directed acyclic graph) DAG is learned based on scoring and search algorithms.
In a preferred embodiment, further comprising:
acquiring scoring data corresponding to the total charging time from an expert database by combining with an expert model;
determining a corresponding weight factor according to the scoring data in combination with a preset corresponding relation table of the scoring data and the weight factor;
and correcting the output result of the Bayesian network model by using the weight factor.
In the embodiment, the weight factor can avoid the correction value from generating larger deviation, and the addition of the weight factor greatly improves the accuracy of the output result of the Bayesian network model and has better reference effect because the weight factor is obtained by combining the expert model.
Referring to fig. 2, a second embodiment of the present invention provides a battery charging dynamic protection apparatus, including:
the charging information acquisition module is used for acquiring charging information of the battery, wherein the charging information comprises residual recyclable time data, charging time data and charging duration information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function;
the charging protection module triggers to generate a first pulse current if the charging voltage of the battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the battery reaches the second charging threshold voltage in the charging process of the battery through a charging power supply; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the battery, and the second pulse current is used for triggering the energy storage device to be coupled with the charging power supply.
In a preferred embodiment, the second charging threshold voltage generation module includes:
the conversion unit is used for generating an overshoot protection threshold value and a fast-rush interval corresponding to the current residual circulatable times based on the ratio of the residual circulatable times to the set total circulatable times, the set overshoot protection threshold value and the fast-rush acceleration interval;
the error correction unit generates a threshold error correction value according to the charging frequency data, the charging time length information of each time and a set loss empirical function;
and the second charging threshold voltage generation unit corrects the overshoot protection threshold value by using the threshold error correction value to obtain the second charging threshold voltage.
In a preferred embodiment, a neural network model module is constructed to build the neural network model.
In a preferred embodiment, the neural network model is a bayesian network model.
In a preferred embodiment, further comprising:
the scoring data acquisition module is used for acquiring scoring data corresponding to the total charging time from the expert database in combination with the expert model;
the weighting factor generation module is used for determining a corresponding weighting factor according to the scoring data in combination with a preset corresponding relation table of the scoring data and the weighting factor;
and the result correction module corrects the output result of the Bayesian network model by using the weight factor.
The relevant effects of the device of the present invention are the same as the corresponding methods described above, and are not described herein in any greater detail.
Referring to fig. 3, a battery charging dynamic protection system according to a third embodiment of the present invention includes: a dynamic battery charging protection device 301 and an automobile charging post 302,
wherein the battery charging dynamic protection device comprises:
the charging information acquisition module is used for acquiring charging information of the automobile lithium battery from an automobile BMS system, wherein the charging information comprises residual recyclable time data, charging time data and charging time information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function;
the charging protection module triggers to generate a first pulse current if the charging voltage of the automobile lithium battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the automobile lithium battery reaches the second charging threshold voltage in the process that the automobile lithium battery is charged through the automobile charging pile; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the automobile lithium battery, and the second pulse current is used for triggering the energy storage device to be coupled with the automobile charging pile.
The relevant effect of the system in the present invention is the same as the above corresponding method, and will not be described herein.
An embodiment of a fourth aspect of the present invention provides a dynamic protection system for charging a lithium battery of an automobile based on an empirical function and a 5G technology, including: a dynamic protection device for charging a battery and an automobile charging pile,
wherein the battery charging dynamic protection device comprises:
the charging information acquisition module is used for acquiring charging information of the automobile lithium battery from an automobile BMS system through the Internet of things technology, wherein the charging information comprises residual recyclable time data, charging time data and charging time length information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function;
the charging protection module triggers to generate a first pulse current if the charging voltage of the automobile lithium battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the automobile lithium battery reaches the second charging threshold voltage in the process that the automobile lithium battery is charged through the automobile charging pile; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the automobile lithium battery, and the second pulse current is used for triggering the energy storage device to be coupled with the automobile charging pile.
According to the automobile lithium battery dynamic charging protection system based on the empirical function and the Internet of things, the charging information is input into a preset neural network model, the neural network model outputs a first charging threshold voltage, a second charging threshold voltage is generated through the charging information and a set loss empirical function, further, when the charging voltage of a battery is in a lower level (lower than the first charging threshold voltage), the charging speed is improved by discharging and supplementing charging through an energy storage device, and when the charging voltage of the battery is in a higher level (higher than the second charging threshold voltage), the charging electric quantity of a charging power supply can be shared by charging through the energy storage device, so that the loss of the battery due to quick charging in the final end charging link is avoided, and the effect of dynamic balance charging is achieved; according to the technical scheme provided by the invention, the first charging threshold voltage and the second charging threshold voltage are dynamically generated by utilizing a neural network model and an empirical function, the first charging threshold voltage and the second charging threshold voltage can be adjusted according to the self-loss condition of the battery, and meanwhile, the output result of the neural network model is corrected through a weighting factor, so that the output result can be dynamically attached to the self-actual condition of each battery, the dynamic protection is more accurate and adaptive, the service life of the battery is greatly prolonged, in addition, the battery information is accurately transmitted in real time by adopting the technology of the internet of things and fusing various wired and wireless networks with the internet, and the charging information is more convenient and intelligent to obtain.
The principle and the implementation mode of the invention are explained by applying specific embodiments in the invention, and the description of the embodiments is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A method for dynamic protection of battery charging, comprising:
acquiring charging information of a battery, wherein the charging information comprises residual recyclable time data, charging time data and charging duration information each time;
inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function; the neural network model is obtained by utilizing the charging information training of different batteries of the same type;
in the process that the battery is charged by a charging power supply, if the charging voltage of the battery is lower than the first charging threshold voltage, triggering to generate a first pulse current, and if the charging voltage of the battery reaches the second charging threshold voltage, triggering to generate a second pulse current; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the battery, and the second pulse current is used for triggering the energy storage device to be coupled with the charging power supply.
2. The battery charging dynamic protection method of claim 1, wherein generating a second charging threshold voltage from the charging information and a set loss empirical function comprises:
based on the ratio of the residual recyclable time data to the set total recyclable time data, an overshoot protection threshold and a fast-rush acceleration interval, generating an overshoot protection threshold and a fast-rush interval corresponding to the current residual recyclable time;
generating a threshold error correction value according to the charging time data, the charging time information of each time and a set loss empirical function;
and correcting the overshoot protection threshold value by using the threshold error correction value to obtain the second charging threshold voltage.
3. The battery charging dynamic protection method of claim 2, further comprising: and establishing the neural network model.
4. The battery charge dynamics protection method of claim 3, wherein the neural network model is a Bayesian network model.
5. The battery charging dynamic protection method of claim 4, further comprising:
acquiring scoring data corresponding to the total charging time from an expert database by combining with an expert model;
determining a corresponding weight factor according to the scoring data in combination with a preset corresponding relation table of the scoring data and the weight factor;
and correcting the output result of the Bayesian network model by using the weight factor.
6. A battery charging dynamic protection device, comprising:
the charging information acquisition module is used for acquiring charging information of the battery, wherein the charging information comprises residual recyclable time data, charging time data and charging duration information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function; the neural network model is obtained by utilizing the charging information training of different batteries of the same type;
the charging protection module triggers to generate a first pulse current if the charging voltage of the battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the battery reaches the second charging threshold voltage in the charging process of the battery through a charging power supply; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the battery, and the second pulse current is used for triggering the energy storage device to be coupled with the charging power supply.
7. The battery charging dynamic protection device of claim 6, further comprising:
and constructing a neural network model module and establishing the neural network model.
8. The battery charging dynamic protection device of claim 7, further comprising:
the scoring data acquisition module is used for acquiring scoring data corresponding to the total charging time from the expert database in combination with the expert model;
the weighting factor generation module is used for determining a corresponding weighting factor according to the scoring data in combination with a preset corresponding relation table of the scoring data and the weighting factor;
and the result correction module corrects the output result of the Bayesian network model by using the weight factor.
9. The utility model provides a car lithium cell dynamic charging protection system which characterized in that includes: battery dynamic protection device and car charging pile, wherein battery charging dynamic protection device includes:
the charging information acquisition module is used for acquiring charging information of the automobile lithium battery from an automobile BMS system, wherein the charging information comprises residual recyclable time data, charging time data and charging time information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function; the neural network model is obtained by utilizing the charging information training of different batteries of the same type;
the charging protection module triggers to generate a first pulse current if the charging voltage of the automobile lithium battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the automobile lithium battery reaches the second charging threshold voltage in the process that the automobile lithium battery is charged through the automobile charging pile; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the automobile lithium battery, and the second pulse current is used for triggering the energy storage device to be coupled with the automobile charging pile.
10. Car lithium cell dynamic charge protection system based on empirical function and thing networking, its characterized in that includes: battery dynamic protection device and car charging pile, wherein battery charging dynamic protection device includes:
the charging information acquisition module is used for acquiring charging information of the automobile lithium battery from an automobile BMS system through the Internet of things technology, wherein the charging information comprises residual recyclable time data, charging time data and charging time length information each time;
the threshold voltage generation module is used for inputting the charging information into a preset neural network model, outputting a first charging threshold voltage by the neural network model, and generating a second charging threshold voltage by the charging information and a set loss empirical function; the neural network model is obtained by utilizing the charging information training of different batteries of the same type;
the charging protection module triggers to generate a first pulse current if the charging voltage of the automobile lithium battery is lower than the first charging threshold voltage, and triggers to generate a second pulse current if the charging voltage of the automobile lithium battery reaches the second charging threshold voltage in the process that the automobile lithium battery is charged through the automobile charging pile; wherein,
the first pulse current is used for triggering an energy storage device to be coupled with the automobile lithium battery, and the second pulse current is used for triggering the energy storage device to be coupled with the automobile charging pile.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110014866.8A CN112865221A (en) | 2021-01-06 | 2021-01-06 | Automobile lithium battery dynamic charging protection system based on experience function and Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110014866.8A CN112865221A (en) | 2021-01-06 | 2021-01-06 | Automobile lithium battery dynamic charging protection system based on experience function and Internet of things |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112865221A true CN112865221A (en) | 2021-05-28 |
Family
ID=76004428
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110014866.8A Pending CN112865221A (en) | 2021-01-06 | 2021-01-06 | Automobile lithium battery dynamic charging protection system based on experience function and Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112865221A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114285060A (en) * | 2021-12-28 | 2022-04-05 | 苏州汇川控制技术有限公司 | Grid-connected charging method, equipment and device for hybrid electric vehicle and readable storage medium |
| CN115366710A (en) * | 2022-10-24 | 2022-11-22 | 沈阳宇龙新能源汽车有限公司 | New energy automobile self-adaptation control system that charges based on big data |
-
2021
- 2021-01-06 CN CN202110014866.8A patent/CN112865221A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114285060A (en) * | 2021-12-28 | 2022-04-05 | 苏州汇川控制技术有限公司 | Grid-connected charging method, equipment and device for hybrid electric vehicle and readable storage medium |
| CN115366710A (en) * | 2022-10-24 | 2022-11-22 | 沈阳宇龙新能源汽车有限公司 | New energy automobile self-adaptation control system that charges based on big data |
| CN115366710B (en) * | 2022-10-24 | 2022-12-27 | 沈阳宇龙新能源汽车有限公司 | New energy automobile self-adaptation control system that charges based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113511082A (en) | Hybrid electric vehicle energy management method based on rule and double-depth Q network | |
| Xu et al. | A soft actor-critic-based energy management strategy for electric vehicles with hybrid energy storage systems | |
| Bo et al. | State of charge estimation based on evolutionary neural network | |
| CN112865221A (en) | Automobile lithium battery dynamic charging protection system based on experience function and Internet of things | |
| CN112215434A (en) | LSTM model generation method, charging duration prediction method and medium | |
| CN111191934A (en) | Multi-target cloud workflow scheduling method based on reinforcement learning strategy | |
| CN112865218A (en) | Automobile lithium battery dynamic charging protection system based on knowledge graph and 5G technology | |
| Ermon et al. | Learning policies for battery usage optimization in electric vehicles | |
| Li et al. | Lithium-ion battery remaining useful life prognostics using data-driven deep learning algorithm | |
| CN112307667A (en) | Method and device for estimating state of charge of storage battery, electronic equipment and storage medium | |
| CN115236526A (en) | Method and device for predicting residual charging time, storage medium and vehicle | |
| CN112803527A (en) | Automobile lithium battery charging dynamic protection system based on experience function and big data | |
| CN112865217A (en) | Neural network and artificial intelligence based automobile lithium battery charging dynamic protection system | |
| Baccari et al. | Deep reinforcement learning for cell balancing in electric vehicles with dynamic reconfigurable batteries | |
| Habib | A comparative study of the machine learning-based energy management system for hydrogen fuel cell electric vehicles | |
| Reddy et al. | A Comparison of the State-of-the-Art Reinforcement Learning Algorithms for Health-Aware Energy & Emissions Management in Zero-emission Ships | |
| CN112865220A (en) | Automobile lithium battery dynamic charging protection system based on neural network and cloud computing | |
| CN112528033A (en) | Knowledge graph multi-hop inference method and device, electronic equipment and storage medium | |
| Shi et al. | The ANN models for SOC/BRC estimation of Li-ion battery | |
| CN112838640A (en) | Automobile lithium battery dynamic charging protection system based on neural network and data fusion | |
| CN112865222A (en) | Automobile lithium battery dynamic charging protection system based on knowledge graph and 5G technology | |
| CN116843500A (en) | Charging station planning method, neural network model training method, device and equipment | |
| CN115236520A (en) | Battery remaining service life prediction method and system based on ISSA-LSTM algorithm | |
| Guo et al. | A lifetime extended energy management strategy for fuel cell hybrid electric vehicles via self-learning fuzzy reinforcement learning | |
| CN118647881A (en) | Battery SOC estimation method and device based on meta-learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210528 |