CN112881930B - Energy storage battery health management prediction method and system based on Internet of things - Google Patents
Energy storage battery health management prediction method and system based on Internet of things Download PDFInfo
- Publication number
- CN112881930B CN112881930B CN202110069586.7A CN202110069586A CN112881930B CN 112881930 B CN112881930 B CN 112881930B CN 202110069586 A CN202110069586 A CN 202110069586A CN 112881930 B CN112881930 B CN 112881930B
- Authority
- CN
- China
- Prior art keywords
- soh
- data
- calculation
- weight
- discharge
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 55
- 230000036541 health Effects 0.000 title claims abstract description 49
- 238000004146 energy storage Methods 0.000 title claims abstract description 23
- 238000004364 calculation method Methods 0.000 claims abstract description 86
- 238000004891 communication Methods 0.000 claims abstract description 23
- 230000003993 interaction Effects 0.000 claims abstract description 23
- 230000008569 process Effects 0.000 claims description 32
- 238000012549 training Methods 0.000 claims description 23
- 238000005070 sampling Methods 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 11
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 9
- 230000005540 biological transmission Effects 0.000 claims description 9
- 229910052744 lithium Inorganic materials 0.000 claims description 9
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000009795 derivation Methods 0.000 claims description 3
- 102220124171 rs146038948 Human genes 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 6
- 239000002253 acid Substances 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000087 stabilizing effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/385—Arrangements for measuring battery or accumulator variables
- G01R31/387—Determining ampere-hour charge capacity or SoC
- G01R31/388—Determining ampere-hour charge capacity or SoC involving voltage measurements
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Secondary Cells (AREA)
Abstract
The invention discloses an energy storage battery health management prediction method and system based on the Internet of things, wherein the prediction method comprises the following steps: step 1: the local BMS module samples and uploads real-time data to the cloud platform through the communication module in combination with the built-in parameters; step 2: the cloud platform acquires data participating in calculation according to the real-time data; step 3: the cloud platform establishes a battery health model; step 4: the cloud platform trains the battery health model through the data sampled for a plurality of times; step 5: the cloud platform obtains an actual SOH value through calculation, and transmits the actual SOH value back to the man-machine interaction module, and the man-machine interaction module displays the actual SOH value.
Description
Technical Field
The invention relates to the field of lithium battery health management, in particular to an energy storage battery health management prediction method and system based on the Internet of things.
Background
Along with the increasing growth of energy storage markets, lithium batteries begin to replace lead-acid batteries due to the characteristics of high energy density, long service life, high bearing power and the like, and meanwhile, the safety of the lithium batteries is a certain gap compared with that of the lead-acid batteries, so that the health management of the batteries becomes an indispensable part of a lithium battery system. In the health management of the battery, the state of health (soh) of the battery is one of the most important two parameters, and directly determines how much power the energy storage system can store, and is closely related to the calculation of the battery capacity soh, so the importance thereof is self-evident. If the state of health and the service life of each battery can be accurately predicted in a complex dynamic use environment, the method has great significance on gradient utilization of the batteries and detection of battery faults, and plays a particularly positive role in the safety of the system.
Currently, there are two general methods for calculating soh, respectively:
first, soh = (ampere-hour integrated charge/open circuit voltage method charge) ×100%. The calculation method must calculate when stabilizing low current to obtain the battery health state relatively close to the true value, and in the actual energy storage project, the working state of high power and continuous change of current is generally adopted. Therefore, it is difficult to reach the condition of calculation soh;
second, the battery state of health is obtained based on the change in the internal resistance of the battery. In this method, the internal resistance of the battery is difficult to visually obtain in the actual energy storage project.
Therefore, how to intuitively and accurately obtain soh of a lithium battery system in an actual energy storage project is a problem to be solved in the industry.
Disclosure of Invention
In order to solve the problems, the invention provides an energy storage battery health management prediction system and method based on the internet of things, which are characterized in that the internet of things platform and a cloud database are trained through a BP neural network algorithm, a temperature, a cut-off voltage, a current, a current voltage, a battery working time, a cycle number, a soc value (state of charge, a percentage of the residual electric quantity of a battery, hereinafter simply referred to as soc), a rated capacity and a relationship between a battery core soh are established, and under the condition that cloud data are more, the invention can enable the battery core soh to approach to a true value more intuitively and accurately to obtain soh of a lithium battery system in an actual energy storage project.
In order to achieve the above purpose, the invention provides an energy storage battery health management prediction method based on the Internet of things, which comprises the following steps:
step 1: the local BMS module samples and uploads real-time data to the cloud platform through the communication module by combining the built-in parameters, wherein the real-time data comprises: current, current soc, current temperature, battery operating time, rated capacity, number of cycles, current voltage, charge cutoff voltage, and discharge cutoff voltage;
step 2: the cloud platform acquires data participating in calculation according to the real-time data, wherein the data participating in calculation comprises discharge multiplying power, overcharge and overdischarge times and discharge depth;
step 3: the cloud platform establishes a battery health model by using the current temperature, the discharge multiplying power, the discharge depth, the cycle times, the overcharging and overdischarging times and the current soc, which are obtained in the step 1 and the step 2, wherein the battery health model comprises an input layer, a hidden layer and an output layer;
step 4: the cloud platform trains the battery health model through the data sampled for a plurality of times;
step 5: the cloud platform obtains an actual SOH value through calculation, and transmits the actual SOH value back to the man-machine interaction module, and the man-machine interaction module displays the actual SOH value.
In an embodiment of the present invention, the method for sampling real-time data in step 1 specifically includes:
obtaining current once every 5 seconds, wherein the current is represented by I, and the unit is A;
acquiring a current SOC every 5 seconds, wherein the current SOC is represented by an SOC;
acquiring the current temperature of the battery every 5 seconds, wherein the current temperature is represented by T, and the unit is the temperature;
recording the battery working Time, wherein the battery working Time is expressed by Time, and the unit is a Hour;
only one rated capacity is required to be obtained, wherein the rated capacity is expressed by Cap, and the unit is AH;
recording the cycle number, wherein the cycle number is expressed by CYC;
acquiring current voltage every 5 seconds, wherein the current voltage is represented by U, and the unit is v;
only one charge cut-off voltage and one discharge cut-off voltage are required to be obtained, the charge cut-off voltage is defined by V over The discharge cut-off voltage is represented by V under The unit of the charge cutoff voltage and the discharge cutoff voltage is v.
In an embodiment of the present invention, the specific method for acquiring the data participating in the calculation in the step 2 includes:
a discharge rate calculated by the formula c=i/Cap, wherein the discharge rate is C;
the number of overcharge and overdischarge times is expressed by Error, if the current voltage U>Cut-off voltage of charge V over When the charge and discharge times are over-charged and over-discharged, the number of times is error+1; if the current voltage U<Cut-off voltage of discharge V under When the charge and discharge times are over-charged and over-discharged, the number of times is error+1; if continuous overcharge and overdischarge exist in the charge and discharge process, the overcharge and overdischarge times Error are accumulated only once;
depth of discharge, expressed in DOD, was obtained by recording the current soc value at which the current I was 0A last time.
In an embodiment of the present invention, in step 3, the input layer, the hidden layer, and the output layer are specifically:
the input layer is a data input part of the battery health model, and the input data comprises the current temperature T, the discharge multiplying power C, the discharge depth DOD, the circulation frequency CYC and the overcharge and overdischarge frequency Error;
a hidden layer comprising a first layer and a second layer, wherein the first layer comprises two neurons, H11 and H12 respectively; the second layer contains two neurons, H21 and H22, respectively;
the output layer, SOH, is indicated by SOH.
In an embodiment of the present invention, a specific calculation process of the battery health model for any one collected data is first data forward transmission calculation, and then reverse transmission calculation error values are specifically:
s31: the specific calculation process from the input layer to the first layer of the hidden layer is as follows:
H11＝T×W TH11 +C×W CH11 +DOD×W DODH11 +CYC×W CYCH11 +Error×W ErrorH11
H12＝T×W TH12 +C×W CH12 +DOD×W DODH12 +CYC×W CYCH12 +Error×W ErrorH12
in which W is TH11 Is the weight between T and H11, W CH11 Weight between C and H11, W DODH11 For weights between DOD and H11, W CYCH11 Is the weight between CYC and H11, W ErrorH11 Weight between Error and H11, W TH12 Is the weight between T and H12, W CH12 Weight between C and H12, W DODH12 For weights between DOD and H12, W CYCH12 Is the weight between CYC and H12, W ErrorH12 Weights between Error and H12;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
H11 out ＝1/(1+e H11 )
H12 out ＝1/(1+e H12 )
wherein H11 out H12 is the output value of H11 out The output value of H12, e is a natural constant;
s32: the specific calculation process from the first layer of the hidden layer to the second layer of the hidden layer is as follows:
H21＝H11 out ×W H11H21 +H12 out ×W H12H21
H22＝H11 out ×W H11H22 +H12 out ×W H12H22
in which W is H11H21 Is the weight between H11 and H21, W H12H21 Is the weight between H12 and H21, W H11H22 Is the weight between H11 and H22, W H12H22 Weights between H12 and H22;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
H21 out ＝1/(1+e H21 )
H22 out ＝1/(1+e H22 )
wherein H21 out For the output value of H21, H22 out An output value of H22;
s33: the specific calculation process from the second layer of the hidden layer to the output layer is as follows:
SOH＝H21 out ×W H21SOH +H22 out ×W H22SOH
in which W is H21SOH Weight between H21 and SOH, W H22SOH Weights between H22 and SOH;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
SOH out ＝1/(1+e SOH )
in SOH out Is the output value of SOH;
s34: the reverse transmission calculation error value is specifically:
E TOTAL ＝0.5×(SOH T -SOH OUT ) 2
wherein E is TOTAL SOH as the total error value T For SOH value in training data, namely factory SOH value stored in cloud database。
In one embodiment of the present invention, step 4 specifically includes:
s41: the cloud platform calls manufacturer data pre-stored in a cloud database;
s42: presetting initial weight values to be 0.1;
s43: training is carried out by continuously repeating the processes from the step 1 to the step 3, and the values of all weights are updated according to each training result and used for the next calculation;
s44: when S43 is repeated 10 ten thousand times, training is stopped, and each weight value at that time is acquired as a weight for final calculation.
In an embodiment of the present invention, the weight calculation in step S43 is obtained by combining the chain rule with the forward derivation formula principle, and is represented by W H21SOH The specific calculation process is as follows:
wherein eta is learning rate and is preset to be 0.5; w (W) H21SOH + For the weights between H21 and SOH at the next sampling calculation,
similarly, W H11H21 The specific calculation process is as follows:
in which W is H11H21 + For the weights between H11 and H21 at the next sampling calculation,
similarly, the same algorithm is used for the calculation of other weights.
In an embodiment of the present invention, the actual SOH value in step 5 is specifically calculated as follows:
s51: step 1 and step 2 are executed again to obtain real-time sampling data of the lithium battery;
s52: and (3) carrying the weight for final calculation obtained through training in the step (4) into the calculation process of the step (3) for calculation to obtain an actual SOH value.
In order to achieve the above purpose, the invention further provides an energy storage battery health management prediction system based on the internet of things, which comprises a local BMS module, a communication module, a man-machine interaction module, a cloud platform and a cloud database, wherein:
the local BMS module is connected with the communication module and is used for collecting battery temperature, real-time current and real-time voltage data and sending the collected data to the communication module, and the local BMS module comprises an RMU module of the commercial energy storage BMS controller;
the man-machine interaction module is connected with the communication module and used for displaying data, and comprises a CMU module of the commercial energy storage BMS controller;
the communication module is also connected with the cloud platform and is used for sending data acquired by the local BMS module to the cloud platform, and the communication module comprises a WIFI and a 100M network card which are carried by the CMU module;
the cloud platform is connected with the man-machine interaction module and is used for displaying the acquired data to the man-machine interaction module after calculation;
the cloud database is connected with the cloud platform and used for storing and calling data.
Compared with the prior art, the invention has the following advantages:
1) By using the cloud to perform data processing and training of the neural network model, more effective data can be obtained, meanwhile, the calculation speed is improved, and the problem that the battery health is affected by various factors but cannot be quantified is solved;
2) The cloud platform is used for carrying out real-time calculation of soh, so that a large amount of measurement time of a BMS (Battery Management System ) end is reduced, the functions of the BMS board are simplified, a complex internal resistance test circuit and a complex internal resistance calculation method are not needed, and the calculation efficiency is high;
3) The generation of accumulated errors is reduced through the training of the training model, and meanwhile, various factors affecting the health of the battery are considered, so that the precision and accuracy of soh are remarkably improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a battery health model according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating weight calculation and updating in a battery health model according to an embodiment of the invention;
FIG. 3 is a system architecture diagram of an embodiment of the present invention;
reference numerals illustrate: 101-a local BMS module; 102-a communication module; 103-a man-machine interaction module; 104-a cloud platform; 105-cloud database.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without any inventive effort, are intended to be within the scope of the invention.
The embodiment of the invention provides an energy storage battery health management prediction method based on the Internet of things, which comprises the following steps of:
step 1: the local BMS module samples and uploads real-time data to the cloud platform through the communication module by combining built-in parameters and the like, wherein the uploaded real-time data comprises: current, current soc, current temperature, battery operating time, rated capacity, number of cycles, current voltage, charge cutoff voltage, discharge cutoff voltage.
The method for sampling the real-time data specifically comprises the following steps:
obtaining current once every 5 seconds, wherein the current is represented by I, and the unit is A;
acquiring a current SOC once every 5 seconds, wherein the current SOC is represented by the SOC;
acquiring the current temperature of the battery every 5 seconds, wherein the current temperature is represented by T, and the unit is DEG C;
recording the battery working Time, wherein the battery working Time is represented by Time, and the unit is a Hour;
only one rated capacity is required to be obtained, the rated capacity is expressed by Cap, and the unit is AH;
recording the circulation times, wherein the circulation times are expressed by CYC;
the current voltage is obtained every 5 seconds, and is represented by U, and the unit is v;
only one charge cut-off voltage and one discharge cut-off voltage are required to be obtained, and the charge cut-off voltage is V over V for discharge cut-off voltage under The unit of the charge cut-off voltage and the discharge cut-off voltage is v.
Step 2: the cloud platform acquires data participating in calculation according to the acquired real-time data, wherein the data participating in calculation comprises discharge multiplying power, overcharge and overdischarge times and discharge depth.
The specific acquisition method of the data participating in the calculation specifically comprises the following steps:
discharge rate, calculated by the formula c=i/Cap, wherein the discharge rate is C;
the number of overcharge and overdischarge times is expressed by Error, if the current voltage U>Cut-off voltage of charge V over When the charge and discharge times are over-charged and over-discharged, the number of times is error+1; if the current voltage U<Cut-off voltage of discharge V under When the charge and discharge times are over-charged and over-discharged, the number of times is error+1; if continuous overcharge and overdischarge exist in the charge and discharge process, the overcharge and overdischarge times Error are accumulated only once；
Depth of discharge, expressed in DOD, was obtained by recording the current soc value at which the current I was 0A last time.
Step 3: the cloud platform establishes a battery health model by using the current temperature T, the discharge multiplying power C, the discharge depth DOD, the circulation frequency CYC, the overcharge and overdischarge frequency Error and the current soc (battery health parameter) obtained in the step 1 and the step 2, wherein the battery health model comprises an input layer, a hidden layer and an output layer.
Fig. 1 is a schematic structural diagram of a battery health model according to an embodiment of the present invention, as shown in fig. 1, wherein in step 3, an input layer, a hidden layer and an output layer are specifically:
the input layer is a data input part of the battery health model, and the input data comprises the current temperature T, the discharge multiplying power C, the discharge depth DOD, the circulation frequency CYC and the overcharge and overdischarge frequency Error;
a hidden layer comprising a first layer and a second layer, wherein the first layer comprises two neurons, H11 and H12 respectively; the second layer contains two neurons, H21 and H22, respectively;
the output layer, SOH, is indicated by SOH.
In this embodiment, the specific calculation process of the battery health model for any one collected data is that data is calculated by forward transmission of the data, and then error values are calculated by reverse transmission, which is specifically as follows:
s31: the specific calculation process from the input layer to the first layer of the hidden layer is as follows:
H11＝T×W TH11 +C×W CH11 +DOD×W DODH11 +CYC×W CYCH11 +Error×W ErrorH11
H12＝T×W TH12 +C×W CH12 +DOD×W DODH12 +CYC×W CYCH12 +Error×W ErrorH12
in which W is TH11 Is the weight between T and H11, W CH11 Weight between C and H11, W DODH11 For weights between DOD and H11, W CYCH11 Is the weight between CYC and H11, W ErrorH11 Weight between Error and H11, W TH12 Is T to H12Weights of each other, W CH12 Weight between C and H12, W DODH12 For weights between DOD and H12, W CYCH12 Is the weight between CYC and H12, W ErrorH12 Weights between Error and H12;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
H11 out ＝1/(1+e H11 )
H12 out ＝1/(1+e H12 )
wherein H11 out H12 is the output value of H11 out An output value of H12, e is a natural constant (its value is about 2.718281828459045);
s32: the specific calculation process from the first layer of the hidden layer to the second layer of the hidden layer is as follows:
H21＝H11 out ×W H11H21 +H12 out ×W H12H21
H22＝H11 out ×W H11H22 +H12 out ×W H12H22
in which W is H11H21 Is the weight between H11 and H21, W H12H21 Is the weight between H12 and H21, W H11H22 Is the weight between H11 and H22, W H12H22 Weights between H12 and H22;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
H21 out ＝1/(1+e H21 )
H22 out ＝1/(1+e H22 )
wherein H21 out For the output value of H21, H22 out An output value of H22;
s33: the specific calculation process from the second layer of the hidden layer to the output layer is as follows:
SOH＝H21 out ×W H21SOH +H22 out ×W H22SOH
in which W is H21SOH Weight between H21 and SOH, W H22SOH Weights between H22 and SOH;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
SOH out ＝1/(1+e SOH )
in SOH out Is the output value of SOH;
s34: the reverse transmission calculation error value is specifically:
E TOTAL ＝0.5×(SOH T -SOH OUT ) 2
wherein E is TOTAL SOH as the total error value T The SOH value in the training data is the factory SOH value stored in the cloud database. Generally, a factory given SOH value is given by a battery manufacturer, and in this embodiment, the factory given SOH value is pre-stored in a cloud database to be used as a SOH value of training data, so as to train a battery health model to obtain a more accurate actual SOH value.
Step 4: the cloud platform trains the battery health model through the data sampled for a plurality of times.
Wherein, step 4 specifically comprises:
s41: the cloud platform calls manufacturer data pre-stored in a cloud database; in this embodiment, factory data pre-stored in the cloud database includes, for example, the foregoing factory SOH values, and factory values of various data used in the calculation process of the present invention.
S42: presetting initial weight values to be 0.1;
s43: training is carried out by continuously repeating the processes from the step 1 to the step 3, and the values of all weights are updated according to each training result and used for the next calculation;
s44: when S43 is repeated for 10 ten thousand times, training is stopped, and each weight value at the moment is obtained and used as the weight for final calculation, and each weight is respectively W H21SOH 、W TH11 、W CH11 、W DODH11 、W CYCH11 、W ErrorH11 、W TH12 、W CH12 、W DODH12 、W CYCH12 、W ErrorH12 、W H11H21 、W H12H21 、W H11H22 、W H12H22 W and W H22SOH 。
In an embodiment of the present invention, each weight in the training process of the battery health model is a dynamic value, each weight in the first sampling calculation of the battery health model is a preset value, each weight in each sampling calculation is calculated and updated according to the value trained by the last sampling from the second sampling, fig. 2 is a schematic diagram of weight calculation and updating in the battery health model according to an embodiment of the present invention, as shown in fig. 2, any weight calculation in step S43 is obtained by combining a chain rule with a forward derivation formula principle, and W is calculated H21SOH The specific calculation process is as follows:
wherein η is a learning rate, and is preset to 0.5 in this embodiment; w (W) H21SOH + For the weights between H21 and SOH at the next sampling calculation,
similarly, W H11H21 The specific calculation process is as follows:
in which W is H11H21 + For the weights between H11 and H21 at the next sampling calculation,
similarly, the calculation process of other weights adopts the same algorithm, and is not repeated here.
Step 5: the cloud platform obtains an actual SOH value through calculation, and transmits the actual SOH value back to the man-machine interaction module, and the man-machine interaction module displays the actual SOH value.
The actual SOH value in step 5 is specifically calculated as follows:
s51: step 1 and step 2 are executed again to obtain real-time sampling data of the lithium battery;
s52: and (3) carrying the weight for final calculation obtained through training in the step (4) into the calculation process of the step (3) for calculation to obtain an actual SOH value.
Fig. 3 is a system architecture diagram of an embodiment of the present invention, as shown in fig. 3, in an embodiment of the present invention, an energy storage battery health management prediction system based on internet of things is provided, which includes a local BMS module (101), a communication module (102), a man-machine interaction module (103), a cloud platform (104), and a cloud database (105), wherein:
the local BMS module (101) is connected with the communication module (102) and is used for collecting battery temperature, real-time current and real-time voltage data and sending the collected data to the communication module (102), wherein the local BMS module (101) comprises an RMU module of a commercial energy storage BMS controller, RMU (Remote Manangement Unit) is a remote monitoring unit, and is hereinafter called RMU for short;
the man-machine interaction module (103) is connected with the communication module (102) and used for displaying data, and the man-machine interaction module (103) comprises a CMU module of a commercial energy storage BMS controller, wherein CMU (Cell monitor Unit) is a single monitoring unit and is called CMU hereinafter;
the communication module (102) is also connected with the cloud platform (104) and is used for sending data acquired by the local BMS module (101) to the cloud platform (104), and the communication module (102) comprises a WIFI and a 100M network card of the CMU module of the man-machine interaction module (103);
the cloud platform (104) is connected with the man-machine interaction module (103) and is used for displaying the acquired data to the man-machine interaction module (103) after calculation;
the cloud database (105) is connected with the cloud platform (104) and is used for storing and calling data.
According to the invention, by using the cloud to perform data processing and training of the neural network model, more effective data can be obtained, meanwhile, the calculation speed is improved, and the problem that the battery health is affected by various factors but cannot be quantified is solved; in addition, the cloud platform is used for carrying out real-time calculation of soh, so that a large amount of measurement time of the BMS end is reduced, the functions of the BMS board are simplified, a complex internal resistance test circuit and a calculation method are not needed, and the calculation efficiency is high; in addition, the generation of accumulated errors is reduced through the training of the training model, and meanwhile, various factors influencing the health of the battery are considered, so that the precision and accuracy of soh are remarkably improved.
Those of ordinary skill in the art will appreciate that: the modules in the apparatus of the embodiments may be distributed in the apparatus of the embodiments according to the description of the embodiments, or may be located in one or more apparatuses different from the present embodiments with corresponding changes. The modules of the above embodiments may be combined into one module, or may be further split into a plurality of sub-modules.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (6)
1. The energy storage battery health management prediction method based on the Internet of things is characterized by comprising the following steps of:
step 1: the local BMS module samples and uploads real-time data to the cloud platform through the communication module by combining the built-in parameters, wherein the real-time data comprises: current, current soc, current temperature T, battery operating time, rated capacity, cycle count CYC, current voltage, charge cutoff voltage, and discharge cutoff voltage;
step 2: the cloud platform acquires data participating in calculation according to the real-time data, wherein the data participating in calculation comprises a discharge multiplying power C, an overcharge and overdischarge frequency Error and a discharge depth DOD;
step 3: the cloud platform establishes a battery health model with the current soc by using the current temperature T, the discharge multiplying power C, the discharge depth DOD, the circulation frequency CYC, the overcharge and overdischarge frequency Error obtained in the step 1 and the step 2, wherein the battery health model comprises an input layer, a hidden layer and an output layer, and the battery health model comprises the following components:
the input layer is a data input part of a battery health model, and the input data comprises the current temperature T, the discharge multiplying power C, the discharge depth DOD, the circulation frequency CYC and the overcharge and overdischarge frequency Error;
the hidden layer comprises a first layer and a second layer, wherein the first layer comprises two neurons, namely H11 and H12; the second layer comprises two neurons, H21 and H22, respectively;
the output layer is SOH and is expressed by SOH;
the battery health model performs forward transmission calculation on any one calculation process of the acquired data, and then performs reverse transmission calculation on an error value;
the specific calculation process of the battery health model for any collected data is as follows:
s31: the specific calculation process from the input layer to the first layer of the hidden layer is as follows:
H11＝T×W TH11 +C×W CH11 +DOD×W DODH11 +CYC×W CYCH11 +Error×W ErrorH11
H12＝T×W TH12 +C×W CH12 +DOD×W DODH12 +CYC×W CYCH12 +Error×W ErrorH12
in which W is TH11 Is the weight between T and H11, W CH11 Weight between C and H11, W DODH11 For weights between DOD and H11, W CYCH11 Is the weight between CYC and H11, W ErrorH11 Weight between Error and H11, W TH12 Is the weight between T and H12, W CH12 Weight between C and H12, W DODH12 For weights between DOD and H12, W CYCH12 Is the weight between CYC and H12, W ErrorH12 Weights between Error and H12;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
H11 out ＝1/(1+e H11 )
H12 out ＝1/(1+e H12 )
wherein H11 out H12 is the output value of H11 out The output value of H12, e is a natural constant;
s32: the specific calculation process from the first layer of the hidden layer to the second layer of the hidden layer is as follows:
H21＝H11 out ×W H11H21 +H12 out ×W H12H21
H22＝H11 out ×W H11H22 +H12 out ×W H12H22
in which W is H11H21 Is the weight between H11 and H21, W H12H21 Is the weight between H12 and H21, W H11H22 Is the weight between H11 and H22, W H12H22 Weights between H12 and H22;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
H21 out ＝1/(1+e H21 )
H22 out ＝1/(1+e H22 )
wherein H21 out For the output value of H21, H22 out An output value of H22;
s33: the specific calculation process from the second layer of the hidden layer to the output layer is as follows:
SOH＝H21 out ×W H21SOH +H22 out ×W H22SOH
in which W is H21SOH Weight between H21 and SOH, W H22SOH Weights between H22 and SOH;
using the activation function y (x) =1/(1+e) x ) And (3) processing data to obtain:
SOH out ＝1/(1+e SOH )
in SOH out Is the output value of SOH;
s34: the reverse transmission calculation error value is specifically:
E TOTAL ＝0.5×(SOH T -SOH OUT ) 2
wherein E is TOTAL SOH as the total error value T SOH values in training data, namely factory SOH values stored in a cloud database;
step 4: the cloud platform trains the battery health model through the data of sampling many times, specifically:
s41: the cloud platform calls manufacturer data pre-stored in a cloud database;
s42: presetting initial weight values to be 0.1;
s43: training is carried out by continuously repeating the processes from the step 1 to the step 3, and the values of all weights are updated according to each training result and used for the next calculation;
s44: when S43 is repeated for 10 ten thousand times, training is stopped, and each weight value at the moment is obtained and used as the weight for final calculation;
step 5: the cloud platform obtains an actual SOH value through calculation, and transmits the actual SOH value back to the man-machine interaction module, and the man-machine interaction module displays the actual SOH value.
2. The method according to claim 1, wherein the method for sampling real-time data in step 1 is specifically:
obtaining current once every 5 seconds, wherein the current is represented by I, and the unit is A;
acquiring a current SOC every 5 seconds, wherein the current SOC is represented by an SOC;
acquiring the current temperature of the battery every 5 seconds, wherein the current temperature is represented by T, and the unit is the temperature;
recording the battery working Time, wherein the battery working Time is expressed by Time, and the unit is a Hour;
only one rated capacity is required to be obtained, wherein the rated capacity is expressed by Cap, and the unit is AH;
recording the circulation times CYC;
acquiring current voltage every 5 seconds, wherein the current voltage is represented by U, and the unit is v;
only one charge cut-off voltage and one discharge cut-off voltage are required to be obtained, the charge cut-off voltage is defined by V over The discharge cut-off voltage is represented by V under Representation ofThe unit of the charge cutoff voltage and the discharge cutoff voltage is v.
3. The method according to claim 1, wherein the specific acquisition method of the data participating in the calculation in the step 2 is:
discharge rate C, calculated by formula c=i/Cap;
the number of overcharge and overdischarge is Error, if the current voltage U>Cut-off voltage of charge V over When the charge and discharge times are over-charged and over-discharged, the number of times is error+1; if the current voltage U<Cut-off voltage of discharge V under When the charge and discharge times are over-charged and over-discharged, the number of times is error+1; if continuous overcharge and overdischarge exist in the charge and discharge process, the overcharge and overdischarge times Error are accumulated only once;
depth of discharge DOD is obtained by recording the current soc value at which the current I was 0A last time.
4. The method according to claim 1, wherein the weight calculation in step S43 is performed by the chain rule in combination with the forward derivation formula principle, in terms of W H21SOH The specific calculation process is as follows:
wherein eta is learning rate and is preset to be 0.5; w (W) H21SOH + For the weights between H21 and SOH at the next sampling calculation,
similarly, W H11H21 The specific calculation process is as follows:
in which W is H11H21 + For the weights between H11 and H21 at the next sampling calculation,
similarly, the same algorithm is used for the calculation of other weights.
5. The method according to claim 4, wherein the actual SOH value specifically calculating process in step 5 is:
s51: step 1 and step 2 are executed again to obtain real-time sampling data of the lithium battery;
s52: and (3) carrying the weight for final calculation obtained through training in the step (4) into the calculation process of the step (3) for calculation to obtain an actual SOH value.
6. The energy storage battery health management prediction system based on the internet of things is used for realizing the method of any one of claims 1 to 5, and is characterized by comprising a local BMS module, a communication module, a man-machine interaction module, a cloud platform and a cloud database, wherein:
the local BMS module is connected with the communication module and is used for collecting battery temperature, real-time current and real-time voltage data and sending the collected data to the communication module, and the local BMS module comprises an RMU module of the commercial energy storage BMS controller;
the man-machine interaction module is connected with the communication module and used for displaying data, and comprises a CMU module of the commercial energy storage BMS controller;
the communication module is also connected with the cloud platform and is used for sending data acquired by the local BMS module to the cloud platform, and the communication module comprises a WIFI and a 100M network card which are carried by the CMU module;
the cloud platform is connected with the man-machine interaction module and is used for displaying the acquired data to the man-machine interaction module after calculation;
the cloud database is connected with the cloud platform and used for storing and calling data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110069586.7A CN112881930B (en) | 2021-01-19 | 2021-01-19 | Energy storage battery health management prediction method and system based on Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110069586.7A CN112881930B (en) | 2021-01-19 | 2021-01-19 | Energy storage battery health management prediction method and system based on Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112881930A CN112881930A (en) | 2021-06-01 |
| CN112881930B true CN112881930B (en) | 2023-07-18 |
Family
ID=76049793
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110069586.7A Active CN112881930B (en) | 2021-01-19 | 2021-01-19 | Energy storage battery health management prediction method and system based on Internet of things |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112881930B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113612269B (en) * | 2021-07-02 | 2023-06-27 | 国网山东省电力公司莱芜供电公司 | Method and system for controlling charge and discharge of battery monomer of lead-acid storage battery energy storage station |
| CN114089190B (en) * | 2021-11-11 | 2023-08-22 | 深圳力高新能技术有限公司 | Battery SOC estimation method based on neural network |
| CN115792627B (en) * | 2022-11-14 | 2023-06-06 | 上海玫克生储能科技有限公司 | SOH analysis and prediction method and device for lithium battery, electronic equipment and storage medium |
| CN116609686B (en) * | 2023-04-18 | 2024-01-05 | 江苏果下科技有限公司 | Battery cell consistency assessment method based on cloud platform big data |
| CN116400227A (en) * | 2023-06-08 | 2023-07-07 | 长安大学 | SOH prediction method, system, equipment and medium for power battery of electric automobile |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111366848A (en) * | 2019-12-31 | 2020-07-03 | 安徽师范大学 | Battery health state prediction method based on PSO-ELM algorithm |
| CN112098845A (en) * | 2020-08-17 | 2020-12-18 | 四川大学 | Lithium battery state estimation method for distributed energy storage system |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110015162A (en) * | 2017-06-30 | 2019-07-16 | 宝沃汽车（中国）有限公司 | Cell health state detection method, device and system and storage medium |
| CN109768589A (en) * | 2018-12-20 | 2019-05-17 | 北京昆兰新能源技术有限公司 | A kind of battery voltage balanced equipment |
| US11054481B2 (en) * | 2019-03-19 | 2021-07-06 | Battelle Energy Alliance, Llc | Multispectral impedance determination under dynamic load conditions |
| CN110386027B (en) * | 2019-06-19 | 2021-03-02 | 东北大学 | Battery management system for electric automobile combining cloud computing and edge computing |
| CN110659722B (en) * | 2019-08-30 | 2023-07-18 | 江苏大学 | Electric vehicle lithium ion battery health state estimation method based on AdaBoost-CBP neural network |
| CN111157897B (en) * | 2019-12-31 | 2022-05-10 | 国网北京市电力公司 | Method and device for evaluating power battery, storage medium and processor |
-
2021
- 2021-01-19 CN CN202110069586.7A patent/CN112881930B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111366848A (en) * | 2019-12-31 | 2020-07-03 | 安徽师范大学 | Battery health state prediction method based on PSO-ELM algorithm |
| CN112098845A (en) * | 2020-08-17 | 2020-12-18 | 四川大学 | Lithium battery state estimation method for distributed energy storage system |
Non-Patent Citations (2)
| Title |
|---|
| 基于卷积神经网络的锂离子电池SOH估算;李超然 等;电工技术学报(19);116-129 * |
| 锂离子电池健康评估和寿命预测综述;刘大同 等;仪器仪表学报(01);5-20 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112881930A (en) | 2021-06-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112881930B (en) | Energy storage battery health management prediction method and system based on Internet of things | |
| CN107957562B (en) | Online prediction method for residual life of lithium ion battery | |
| CN110568361B (en) | Method for predicting health state of power battery | |
| US11703548B2 (en) | Methods and systems for accelerated determining of state of health using incremental capacity analysis | |
| Kim et al. | Online SOC and SOH estimation for multicell lithium-ion batteries based on an adaptive hybrid battery model and sliding-mode observer | |
| Azis et al. | State of charge (SoC) and state of health (SoH) estimation of lithium-ion battery using dual extended kalman filter based on polynomial battery model | |
| CN105425154B (en) | A kind of method of the state-of-charge for the power battery pack for estimating electric automobile | |
| CN113777501B (en) | SOH estimation method of battery module | |
| CN107169170B (en) | A kind of prediction technique of battery remaining power | |
| FR3118310A1 (en) | METHOD FOR ESTIMATING THE LIFETIME OF AN ENERGY STORAGE SYSTEM | |
| Zhu et al. | The SOH estimation of LiFePO4 battery based on internal resistance with Grey Markov Chain | |
| WO2022144542A1 (en) | Method for predicting condition parameter degradation of a cell | |
| Lu et al. | Modeling discharge characteristics for predicting battery remaining life | |
| Liu et al. | A new dynamic SOH estimation of lead‐acid battery for substation application | |
| CN114114051B (en) | Battery aging value determining method and device and electronic equipment | |
| CN118050643A (en) | Ocv estimation method and system based on curve fitting and electronic equipment | |
| Wu et al. | State-of-charge and state-of-health estimating method for lithium-ion batteries | |
| Chaoui et al. | State of charge estimation of LiFePO 4 batteries with temperature variations using neural networks | |
| Jiani et al. | Li-ion battery SOC estimation using EKF based on a model proposed by extreme learning machine | |
| Kim et al. | Design of State of Charge and Health Estimation for Li-ion Battery Management System | |
| Han et al. | A new SOH prediction model for lithium-ion battery for electric vehicles | |
| CN106646260A (en) | SOC estimation method for BMS system based on genetic neural network | |
| CN113391225B (en) | Lithium battery state-of-charge estimation method considering capacity degradation | |
| CN114764124A (en) | Lithium battery SOC estimation method based on GAN and LSTM | |
| Li et al. | Battery State of Charge Probabilistic Estimation Using Natural Gradient Boosting |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: Room 3041, 3rd Floor, No. 6 Chuangye Road, Shangdi Information Industry Base, Haidian District, Beijing 100085 Patentee after: Beijing Kinglong New Energy Technology Co.,Ltd. Country or region after: China Address before: 100085 Room 308, 3 / F, building 1, No. 11, information road, Haidian District, Beijing Patentee before: Beijing Kinglong New Energy Technology Co.,Ltd. Country or region before: China |