CN102324582A - Intelligent maintenance device of multifunctional lead-acid battery and capacity prediction method - Google Patents
Intelligent maintenance device of multifunctional lead-acid battery and capacity prediction method Download PDFInfo
- Publication number
- CN102324582A CN102324582A CN201110231278A CN201110231278A CN102324582A CN 102324582 A CN102324582 A CN 102324582A CN 201110231278 A CN201110231278 A CN 201110231278A CN 201110231278 A CN201110231278 A CN 201110231278A CN 102324582 A CN102324582 A CN 102324582A
- Authority
- CN
- China
- Prior art keywords
- battery
- storage battery
- capacity
- output
- network
- 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.)
- Granted
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02E60/10—Energy storage using batteries
Landscapes
- Secondary Cells (AREA)
Abstract
Relating to maintenance technologies of lead-acid batteries, the invention provides an intelligent maintenance device of a multifunctional lead-acid battery and a capacity prediction method. The intelligent maintenance device of a multifunctional lead-acid battery in the invention is used for on-line test of lead-acid battery capacity, automatic constant current charge and discharge, as well as capacity recovery of a degraded battery. The device comprises a charging unit, a discharging unit, a master control module and a measurement control circuit. The provided lead-acid battery capacity prediction algorithm through BP neural network optimization based on genetic annealing can be realized in the control module and an upper computer, and the capacity prediction error is smaller than a 15% prediction error of present domestic battery capacity rapid testers. The device of the invention can be used in on-line health status test of lead-acid battery capacity and can realize capacity recovery of a degraded battery, as well as extend the battery service life.
Description
Technical field
The present invention relates to a kind of intelligent maintenance device and capacity predict method of multi-functional lead acid accumulator, be a kind ofly carry out that the lead acid accumulator automatic constant current discharges and recharges, the intelligent maintenance device of capacity predict and deterioration capacity resuming.
Background technology
Lead acid accumulator is a kind of energy-storage travelling wave tube of existing more than 100 year applicating history, since low price, stable performance, and technology and manufacturing process are ripe, are widely used in a plurality of applications such as electric power, communication, automobile, electric motor car.Compare with the common lead acid accumulator; Valve-control sealed lead acid battery is owing to adopt inner oxygen complex technique; Alleviated the loss of electrolyte greatly; Thereby make storage battery long service under non-maintaining state, and have advantages such as volume is little, explosion-proof, voltage is stable, pollution-free, in light weight, discharge performance is high, maintenance is little.Yet owing to the reasons such as design, production technology and working service of storage battery itself, the battery early failure phenomena often has generation, especially homemade lead acid accumulator, and what have can only use 2~3 years, was significantly shorter than life expectancy.Because batteries is to be composed in series by each cell; The capacity of batteries is by the minimum monomer decision of capacity in the whole group; As long as there is any backward cell in the batteries; The capacity that makes whole Battery pack is seriously descended, directly cause the effective time of batteries to reduce.To present actual conditions; Utilize modern batteries measuring technique, the information processing technology, power electronic technology and microprocessor control technology; Develop the multi-functional lead acid accumulator intelligent maintenance equipment that a kind of ability is tested lead acid accumulator health status fast, accurately, safely, online and realized the deterioration capacity resuming; To prolonging the useful life of lead acid accumulator, reduce the failure rate of power-supply system, crucial meaning is arranged.
Lead acid accumulator is the electrochemical system of a complicacy, and the residual capacity of storage battery receives multiple factor affecting such as temperature, discharging current, cell degradation, and it is carried out volume test fast and accurately is very difficult.Because the importance of battery capacity prediction, scholar both domestic and external never stops for the exploration of the capacity predict method of high accuracy and high applicability.Some current main method have: (1) battery voltage measuring method.(2) check electric discharge.(3) electricity is led (internal resistance) mensuration.Because the relation between the external parameter of lead acid accumulator and the capacity presents the non-linear of complicacy, these methods all in all effect are unsatisfactory.Except said method, also have a kind of capacity predict method that is based on the storage battery equivalent-circuit model, this model method can be divided into two big types: one type is empirical model, like the Peukert equation, open circuit voltage-ampere-hour model etc.Another kind of is equivalent-circuit model, like calorifics-electrodynamic model, quadravalence dynamic model; Shepard model and Martin model combination or the like, empirical model be the scientific research personnel through the model that a large amount of experimental summaries come out, be characterized in simple in structure; Amount of calculation is few, on engineering, is widely used.But the problem of considering owing to each empirical model emphasizes particularly on different fields a little, and is not comprehensive, belongs to constant current charge-discharge basically; Deficiency is also arranged in the practical application simultaneously: the one, because model order is higher, calculate relatively difficulty; The 2nd, the foundation of battery model need be confirmed a considerable amount of empirical parameters, and this brings bigger trouble for application person; It is low that a series of like this deficiency has caused current domestic and international capacity of lead acid battery to test intelligent degree, and the capacity predict error is big.
Summary of the invention
The object of the present invention is to provide a kind of intelligent maintenance device and capacity predict method of multi-functional lead acid accumulator; This device provides a kind of capacity of lead acid battery on-line testing that has; The multi-functional lead acid accumulator intelligent maintenance equipment that automatic constant current discharges and recharges and carries out the recovery of deterioration battery capacity; The capacity of lead acid battery Forecasting Methodology can solve because of the outer parameter of lead acid accumulator causes and be complicated non-linear relation with battery itself and cause the big problem of capacity of lead acid battery predicated error, has the high and high characteristics of intelligent degree of precision of prediction.
To achieve these goals, the present invention adopts following technical scheme:
A kind of multi-functional lead acid accumulator intelligent maintenance device; Comprise charhing unit, main control module, discharge cell and circuit of measurement and control; Wherein: aforementioned charhing unit comprises input rectifying filter circuit and switch power module, and the input rectifying filter circuit is connected with switch power module.
Aforementioned main control module comprises that single-chip microcomputer, man-machine interface, communication interface, DAC charging control unit, I/O interface, PWM isolate output unit and ADC measuring unit, and single-chip microcomputer is isolated output unit and is connected with the ADC measuring unit with man-machine interface, communication interface, DAC charging control unit, I/O interface, PWM respectively.
Aforementioned circuit of measurement and control comprises output dip switch A, battery voltage detection battery reversal connection detection, output dip switch B, current detecting and temperature detection; Said battery voltage detection battery reversal connection detects the positive and negative polarities that input connects storage battery, and output connects the input of ADC measuring unit; After current detecting is gathered the current signal of storage battery, the input of input ADC measuring unit; After temperature detection is gathered the temperature of PTC module, the input of input ADC measuring unit; Aforementioned output dip switch A one end is connected with the positive and negative polarities of storage battery, and the other end is connected with switch power module; Aforementioned output dip switch B one end is connected with the positive and negative polarities of storage battery, and the other end is connected with the PTC module.
The output of aforementioned DAC charging control unit is connected with the input of switch power module; The output that aforementioned PWM isolates output unit is connected with the input of PTC driver element; Aforementioned I/O interface is connected with switch power module with cooling fan, output dip switch A, output dip switch B simultaneously, realizes the I/O control between single-chip microcomputer and each module.
Aforementioned PTC module is made up of the parallel connection of some PTC thermistors; Aforementioned currents detects the discharging current of measuring storage battery in real time, and through the single-chip microcomputer operational analysis, real-time regulated PWM isolates the duty ratio of the pwm signal of output unit output, after the PCT driver module drives, accomplishes constant-current discharge through power switch control PTC module.
What the battery capacity prediction operational analysis of aforementioned single-chip microcomputer comprised thes contents are as follows:
(1) foundation of BP neural net
Network configuration is divided into three layers: input layer, hidden layer and output layer, input layer number are 2: discharging current of storage battery
and accumulator voltage
; The implicit number of plies is 1, comprises 6 latent layer unit; Output layer node number is 1: the residual capacity of storage battery
; Hidden node and output node all adopt the Sigmoid function as excitation function, and the training algorithm of network adopts momentum gradient decline back-propagation algorithm.
(2) adopt genetic Annealing algorithm optimization network weight
A) the basic operation parameter of initialization population and network comprises: population scale, the maximum algebraically of evolving, crossover probability, annealing initial temperature
, temperature damping's function and MapkoB chain length;
B) calculate each individual fitness of evaluation;
C) judge whether to satisfy the optimization criterion, then get into step (3) if satisfy, otherwise continue next step;
D) adopt that corresponding genetic operation operator is selected, intersection and mutation operation;
E) optimization of annealing is returned step b) then and is continued circulation.
(3) image data is trained network
After accomplishing abovementioned steps (2); Utilize neural net to have self study character; Discharging current that collects storage battery
and accumulator voltage
are applied in the training of BP neural net, and training process is following:
F) the maximum frequency of training and the minimal error of network are set;
G) with the discharging current
of storage battery and accumulator voltage
training sample as network; Fan-in network adopts momentum gradient decline back-propagation algorithm to train, up to smaller or equal to minimal error or till reaching maximum frequency of training;
Said apparatus adopts the BP neural network algorithm of optimizing based on genetic Annealing as the capacity of lead acid battery Forecasting Methodology, and step is following:
(1) foundation of BP neural net
Network configuration is divided into three layers: input layer, hidden layer and output layer, input layer number are 2: discharging current of storage battery
and accumulator voltage
; The implicit number of plies is 1, comprises 6 latent layer unit; Output layer node number is 1: the residual capacity of storage battery
; Hidden node and output node all adopt the Sigmoid function as excitation function, and the training algorithm of network adopts momentum gradient decline back-propagation algorithm.
(2) adopt genetic Annealing algorithm optimization network weight
A) the basic operation parameter of initialization population and network comprises: population scale, the maximum algebraically of evolving, crossover probability, annealing initial temperature
, temperature damping's function and MapkoB chain length;
B) calculate each individual fitness of evaluation;
C) judge whether to satisfy the optimization criterion, then get into step (3) if satisfy, otherwise continue next step;
D) adopt that corresponding genetic operation operator is selected, intersection and mutation operation;
E) optimization of annealing is returned step b) then and is continued circulation.
(3) image data is trained network
After accomplishing abovementioned steps (2); Utilize neural net to have self study character; Discharging current that collects storage battery
and accumulator voltage
are applied in the training to the BP neural net, and training process is following:
F) the maximum frequency of training and the minimal error of network are set;
G) with the discharging current
of storage battery and accumulator voltage
training sample as network; Fan-in network adopts momentum gradient decline back-propagation algorithm to train, up to smaller or equal to minimal error or till reaching maximum frequency of training;
G in the abovementioned steps (3)) said accumulator voltage
is detected by the reversal connection of battery voltage detection battery and collects.
D in the abovementioned steps (2)) said mutation operation adopts the self adaptation aberration rate, and the variation probability is:
In the formula:
representes current evolutionary generation;
expression is extremely current on behalf of the algebraically that only takes place continuously to evolve since evolving last time, and
is the coefficient of aberration rate raising.
The present invention has the following advantages:
(1) main control module adopts 16 dsPIC30F series monolithics, realizes parameter setting, operation control, battery capacity forecast analysis and data storage management to system.
(2) main circuit of charge switch power module adopts phase shifting control full-bridge ZVT-PWM soft switch power converter technique; Reduced the switching process loss; Switching frequency can reach 200KHz, effectively reduces the volume and weight of transformer, has improved the conversion efficiency of charge power supply; And the excessive problem of switch stress can not take place, guarantee that system safety is reliable.
(3) device adopts modularity design technology; Technology such as comprehensive utilization intelligent predicting, power electronics and microprocessor control; Have the battery capacity fast prediction, automatic constant current discharges and recharges and multiple function such as deterioration capacity resuming; Can prolong the useful life of storage battery, reduce the failure rate of power-supply system.
(4) adopt the intelligent capacity predict algorithm of optimizing the BP neural net based on genetic Annealing, capacity fast prediction error is less than 10%, and precision of prediction can not reduce along with the use of storage battery, is superior to the predicated error of domestic existing apparatus 15%.
Description of drawings
Fig. 1 is equipments overall structure figure of the present invention;
Fig. 2 the present invention is based on genetic Annealing to optimize BP network modelling flow chart;
Fig. 3 is a lead acid accumulator residual capacity prediction model test result scattergram of the present invention;
Among Fig. 1, the 1-charhing unit; The 2-main control module; The 3-discharge cell; The 4-circuit of measurement and control; 1.1-input rectifying filter circuit; 1.2-switch power module; 2.1-single-chip microcomputer; 2.2-man-machine interface; 2.3-communication interface; 2.4-DAC charging control unit; 2.5-I/O interface; 2.6-PWM isolates output unit; 2.7-ADC measuring unit; 3.1-PTC driver element; 3.2-power switch; 3.3-PTC module; 3.4-cooling fan; 4.1-output dip switch A; 4.2-the reversal connection of battery voltage detection battery detects; 4.3-output dip switch B; 4.4-current detecting; 4.5-temperature detection.
Embodiment
Further specify below in conjunction with accompanying drawing.
Referring to Fig. 1, a kind of multi-functional lead acid accumulator intelligent maintenance device comprises charhing unit 1, main control module 2, discharge cell 3 and circuit of measurement and control 4, wherein:
Aforementioned charhing unit 1 comprises input rectifying filter circuit 1.1 and switch power module 1.2 for lead-acid batteries provides charge power supply, and input rectifying filter circuit 1.1 is connected with switch power module 1.2.
After the correct connection of batteries; Output dip switch B4.3 breaks off; The 220V alternating current converts the high direct voltage of 310V to behind input rectifying filter circuit 1.1, again through DC/DC switch power module 1.2; Converting the controlled direct current about 14.4V to, is storage battery power supply through output dip switch A4.1.In the input voltage normal range (NR); When launching charge function; Input gets into switch power module 1.2 behind current rectifying and wave filtering circuit 1.1; Main control module 2 is through the start and stop and the output voltage of I/O interface 2.5 and DAC charging control unit 2.4 control switching modules 1.2, realizes functions such as the common charging of batteries and fast-pulse chargings.Before charhing unit is the batteries power supply; Carry out the batteries reversal connection and detect, in case batteries polarity connects instead or accumulator battery voltage (showing the bad or batteries open circuit damage of contact) on the low side; Output dip switch A4.1 is not closed, is not battery charging.In charging process, detect the terminal voltage and the charging current of batteries in real time, in case break down, stop charging immediately; Whether charging capacity and the charging termination condition of judging batteries simultaneously occur, and stop charging when satisfying condition.
Aforementioned main control module 2 comprises that single-chip microcomputer 2.1, man-machine interface 2.2, communication interface 2.3, DAC charging control unit 2.4, I/O interface 2.5, PWM isolate output unit 2.6 and ADC measuring unit 2.7, and single-chip microcomputer 2.1 is isolated output unit 2.6 and is connected with ADC measuring unit 2.7 with man-machine interface 2.2, communication interface 2.3, DAC charging control unit 2.4, I/O interface 2.5, PWM respectively.
Single-chip microcomputer 2.1 is 16 dsPIC30F series monolithics.Man-machine interface 2.2 is made up of 4 cun Chinese graphic lcds and 6 buttons.The ADC measuring unit 2.7 of main control module 2 uses the ADC passage of single-chip microcomputer 2.1 inside.The DAC charging control unit realizes that through single-chip microcomputer expansion D/A converter communication interface is RS-232C.Main control module 2 is realized parameter setting, the operation control to system; Also be responsible for coordinating the work of charhing unit 1 and discharge cell 3; Simultaneously the discharging current
and battery terminal voltage
utilization of the storage battery that collects are calculated based on genetic Annealing optimization BP neural net; Draw the residual capacity
of storage battery, and realize that the repeatedly little current cycle charge/discharge capacity of deterioration storage battery recovers.Communication interface 2.3 in the main control module 2 can be connected to upper PC, realizes measurement, control and the capacity predict of lead acid accumulator.The software platform of main control module 2 uses the uC/OS-II embedded OS, has improved the information processing capability of control unit.
Aforementioned discharge cell 3 provides the constant-current discharge load for batteries; Comprise PTC driver element 3.1, power switch 3.2, PTC module 3.3 and cooling fan 3.4; PTC driver element 3.1 outputs connect the input of power switch 3.2, and the output of power switch 3.2 connects the input of PTC module 3.3.After the correct connection of batteries; Output dip switch A4.1 breaks off, and output dip switch B4.3 is closed, and the PWM of main control module 2 isolates output unit 2.6 and produces the PWM modulation signal; Through obtaining the pwm signal that multichannel is isolated after optocoupler 6N137 and the multichannel level translator CD4050 conversion; Behind the multi-channel PWM signal input PTC driver module 3.1,, realize constant-current discharge through the multi-disc PTC thermistor of power switch 3.2 control PTC modules 3.3.I/O interface 2.5 output control cooling fans 3.4 operations of main control module 2 are outside the heat discharger with the PTC generation.The ADC measuring unit 2.7 of main control module 2 is through the temperature of temperature detection 4.5 measurement PTC modules 3.3, in case temperature is too high, then main control module 2 is reported to the police and protected processing.
Aforementioned circuit of measurement and control 4 comprises output dip switch A4.1, battery voltage detection battery reversal connection detection 4.2, output dip switch B4.3, current detecting 4.4 and temperature detection 4.5; The reversal connection of aforementioned battery voltage detection battery detects the positive and negative polarities that 4.2 inputs connect storage battery, and output connects the input of ADC measuring unit 2.7; After current detecting 4.4 is gathered the current signal of storage battery, the input of input ADC measuring unit 2.7; After temperature detection 4.5 is gathered the temperature of PTC module 3.3, the input of input ADC measuring unit 2.7; Aforementioned output dip switch A4.1 one end is connected with the positive and negative polarities of storage battery, and the other end is connected with switch power module 1.2; Aforementioned output dip switch B4.3 one end is connected with the positive and negative polarities of storage battery, and the other end is connected with PTC module 3.3.
The output of aforementioned DAC charging control unit 2.4 is connected with the input of switch power module 1.2; The output that aforementioned PWM isolates output unit 2.6 is connected with the input of PTC driver element 3.1; Aforementioned I/O interface 2.5 is connected with cooling fan 3.4, output dip switch A4.1, output dip switch B4.3 and switch power module 1.2 simultaneously, realizes the I/O control between single-chip microcomputer 2.1 and each module.
Aforementioned PTC module 3.3 is made up of the parallel connection of some PTC thermistors; Aforementioned currents detects 4.4 and detects the battery discharging electric current in real time; Through single-chip microcomputer 2.1 operational analyses; Real-time regulated PWM isolates the duty ratio of the pwm signal of output unit 2.6 outputs, after PTC driver element 3.1 drives, through power switch 3.2 control PTC modules 3.3, accomplishes constant-current discharge.
The co-ordination of aforementioned main control module 2 control charhing units 1 and discharge cell 3 adopts repeatedly little electric current constant current charge-discharge to realize the capacity restoration of deterioration lead acid accumulator.
The concrete steps of its capacity restoration are:
A) constant current charge.With the little electric current constant current charge of 10 hour rates, very fast if charging voltage rises, the violent gassing of storage battery explains that the sulfuration of storage battery is more serious, then should reduce charging current earlier, and available 20 hour rates or the charging of 30 hour rates are charged for a long time; The initial no current if charge then improves charging voltage and charges, and until there being electric current to produce, transfers the little electric current constant current charge of 20 hour rates or 30 hour rates again to.
B) constant-current discharge.The little electric current constant-current discharge of 10 hour rates is to U=10.8V/ only (the storage battery rated voltage is 12V), calculates discharge capacity (C=i * t).
C) constant current charge.Judge the degree that storage battery recovers according to discharge capacity, determine this charging current value.Less like discharge capacity, can continue with little electric current constant current charge; If discharge capacity is bigger, explain that the battery capacity recovery effects is better, then available 10 hour rates charging.
D) constant-current discharge.Method is identical with step b).
E) repeating step c repeatedly), d), reach or near rated capacity until the discharge capacity of storage battery.
Said apparatus adopts the BP neural network algorithm optimized based on the genetic Annealing capacity predict method as lead acid accumulator, and step is following:
(1) foundation of BP neural net
Network configuration is divided into three layers: input layer, hidden layer and output layer, input layer number are 2: discharging current of storage battery
and accumulator voltage
; The implicit number of plies is 1, comprises 6 latent layer unit, has 25 weights (containing 7 threshold values), and the weights scope is [10.0 10.0]; Output layer node number is 1: the residual capacity of storage battery
; Hidden node and output node all adopt the Sigmoid function as excitation function, and the training algorithm of network adopts momentum gradient decline back-propagation algorithm.
(2) adopt genetic Annealing algorithm optimization network weight
Referring to Fig. 2, the whole flow process of capacity of lead acid battery prediction algorithm specifically is divided into the two large divisions, and wherein the first step is optimized the process of network weight for adopting genetic Annealing, and detailed process is following:
A) the basic operation parameter of initialization population and network comprises: population scale, the maximum algebraically of evolving, crossover probability, annealing initial temperature
, temperature damping's function and MapkoB chain length;
B) calculate each individual fitness of evaluation;
C) judge whether to satisfy the optimization criterion, then get into step (3) if satisfy, otherwise continue next step;
D) adopt that corresponding genetic operation operator is selected, intersection and mutation operation;
E) optimization of annealing is returned step b) then and is continued circulation.
Wherein, fitness function: be taken as inverse, that is: about the root-mean-square error of weights
In the formula;
is the actual output vector of network;
is the target vector of network, and
is number of training.
Select operator: the individuality in each generation confirms to produce individual father's chromosome of future generation according to the fitness rule of three, and adopts optimum reserved strategy, and the optimum individual in the per generation colony is directly remained in the population of future generation.
Crossover operator: select two parent vectors
and
to intersect with certain crossover probability
, press following formula and produce a daughter
as parent:
In the formula;
is confirmable or the value of picked at random;
,
is the binary number of picked at random.In the iterative process of reality, keep the parent vector after the interlace operation at random and abandon another parent vector.
Mutation operator: it is individual for
to establish a certain parent; By
to
when carrying out mutation operation; If the excursion of change point
is [
], then new genic value
is:
In the formula;
representes evolutionary generation;
is the binary number of picked at random;
is the adjustment function of variation; It representes [0
] meet the random number of non-uniform Distribution in the scope; Be tending towards 0 gradually along with the increase of evolutionary generation
, be defined as:
In the formula;
is the random number in [0 1];
is system parameters; Value is 2.0, and
is maximum evolutionary generation.
Mutation operation adopts the self adaptation aberration rate, and the variation probability is:
Wherein,
representes current evolutionary generation;
expression is extremely current on behalf of the algebraically that only takes place continuously to evolve since evolving last time; The coefficient that
improves for aberration rate, value is 0.005.
(3) image data is trained network
Behind completing steps (2); Utilize neural net to have self study character; The present invention is applied to the discharging current
and the accumulator voltage
of the storage battery that collects in the training of BP neural net, and training process is following:
F) the maximum frequency of training epochs=8000 and minimal error error=1.4 * 10 of network are set
-4
G) will carry out constant-current discharge with 5A, 10A respectively to the sealed lead-acid accumulator free from maintenance of 12V/100Ah, the discharge off condition is as the criterion with the battery discharging final voltage 10.8V of storage battery factory man regulation.And utilize said apparatus that the discharging current
and the accumulator voltage
of the storage battery under 5A, the 10A discharge condition are gathered; With its training sample fan-in network as network; Train with momentum gradient decline back-propagation algorithm, up to smaller or equal to minimal error or till reaching maximum frequency of training.
Abovementioned steps; (3) g in) discharging current
and the accumulator voltage
of aforementioned storage battery are to be detected by aforementioned currents; (4.4) and the reversal connection of battery voltage detection battery detect; (4.2) detected.
When said apparatus is gathered the discharging current
of the storage battery under 5A, the 10A discharge condition and accumulator voltage
; Record
and
; Adopt ampere-hour method record to put capacity, and conversion is the residual capacity C of storage battery according to the Peukert equation.The residual capacity
of the storage battery of exporting behind this value and the network training is compared.
Aforesaid genetic Annealing algorithm is a kind of algorithm that genetic algorithm and simulated annealing are combined.The parameter of genetic algorithm is set in this test: population scale n=20; Crossover probability
=0.5, variation probability
=0.1; Evolutionary generation is made as 100; The annealing algorithm parameter is set to: temperature decline index is 0.95; The parameter of BP algorithm is set to: learning rate
=0.5, momentum constant
=0.5.Sealed lead-acid accumulator free from maintenance to certain 12V/100Ah under the free position carries out constant-current discharge with 5A, 10A respectively, under every kind of discharging current, respectively gets 10 groups of test sample book data, and is as shown in table 1; Adopt the residual capacity
of the storage battery that obtains behind the above-mentioned network training and be that the residual capacity C of storage battery is more as shown in table 2 by Peukert equation conversion.
Table 1 network training data
Table 2 capacity of lead acid battery prediction contrast
Test result according to table 2 shows, uses device of the present invention and battery capacity Forecasting Methodology that the capacity of storage battery is predicted, its maximum relative error is 8.44%, and minimal error is 0.89%, respectively corresponding the 13rd group and the 1st group of data.Predicated error is superior to the predicated error of domestic existing apparatus 15% all below 10%.
The present invention can also train network through gathering dissimilar storage battery data, obtains the capacity predict model of multiple storage battery, thereby can predict the residual capacity of multiple storage battery.
Claims (6)
1. multi-functional lead acid accumulator intelligent maintenance device is characterized in that: comprise charhing unit (1), main control module (2), discharge cell (3) and circuit of measurement and control (4), wherein:
The input of the switch power module (1.2) in the said charhing unit (1) is connected with DAC charging control unit (2.4) in the main control module (2);
Said circuit of measurement and control (4) comprises that output dip switch A (4.1), the reversal connection of battery voltage detection battery detect (4.2), output dip switch B (4.3), current detecting (4.4) and temperature detection (4.5); Said battery voltage detection battery reversal connection detects the positive and negative polarities that (4.2) input connects storage battery, and output connects the input of the ADC measuring unit (2.7) in the main control module (2); After current detecting (4.4) is gathered the current signal of storage battery, import the input of said ADC measuring unit (2.7); After temperature detection (4.5) is gathered the temperature of the PTC module (3.3) in the discharge cell (3), import the input of said ADC measuring unit (2.7); Said output dip switch A (4.1) one ends are connected with the positive and negative polarities of storage battery, and the other end is connected with said switch power module (1.2); Said output dip switch B (4.3) one ends are connected with the positive and negative polarities of storage battery, and the other end is connected with said PTC module (3.3);
I/O interface (2.5) in the said main control module (2) respectively with discharge cell (3) in cooling fan (3.4), said switch power module (1.2), said output dip switch A (4.1) and output dip switch B (4.3) be connected;
Said current detecting (4.4) detects the discharging current
of storage battery in real time; The battery capacity prediction operational analysis of the single-chip microcomputer (2.1) in said main control module (2); Real-time regulated PWM isolates the duty ratio of the pwm signal of output unit (2.6) output, after PTC driver module (3.1) drives, accomplishes constant-current discharge through power switch (3.2) control PTC module (3.3);
What the battery capacity prediction operational analysis of said single-chip microcomputer (2.1) comprised thes contents are as follows:
(1) foundation of BP neural net
Network configuration is divided into three layers: input layer, hidden layer and output layer, input layer number are 2: discharging current of storage battery
and accumulator voltage
; The implicit number of plies is 1, comprises 6 latent layer unit; Output layer node number is 1: the residual capacity of storage battery
; Hidden node and output node all adopt the Sigmoid function as excitation function, and the training algorithm of network adopts momentum gradient decline back-propagation algorithm;
(2) adopt genetic Annealing algorithm optimization network weight
A) the basic operation parameter of initialization population and network comprises: population scale, the maximum algebraically of evolving, crossover probability, annealing initial temperature
, temperature damping's function and MapkoB chain length;
B) calculate each individual fitness of evaluation;
C) judge whether to satisfy the optimization criterion, then get into step (3) if satisfy, otherwise continue next step;
D) adopt that corresponding genetic operation operator is selected, intersection and mutation operation;
E) optimization of annealing is returned step b) then and is continued circulation;
(3) image data is trained network
Behind completing steps (2); Utilize neural net to have self study character; Discharging current that collects storage battery
and accumulator voltage
are applied in the training of BP neural net, and training process is following:
F) the maximum frequency of training and the minimal error of network are set;
G) with the discharging current
of storage battery and accumulator voltage
training sample as network; Fan-in network adopts momentum gradient decline back-propagation algorithm to train, up to smaller or equal to minimal error or till reaching maximum frequency of training;
2. according to the said a kind of multi-functional lead acid accumulator intelligent maintenance device of claim 1; It is characterized in that: said charhing unit (1) comprises input rectifying filter circuit (1.1) and switch power module (1.2), and input rectifying filter circuit (1.1) is connected with switch power module (1.2);
Said main control module (2) comprises that single-chip microcomputer (2.1), man-machine interface (2.2), communication interface (2.3), DAC charging control unit (2.4), I/O interface (2.5), PWM isolate output unit (2.6) and ADC measuring unit (2.7), and single-chip microcomputer (2.1) is isolated output unit (2.6) and is connected with ADC measuring unit (2.7) with man-machine interface (2.2), communication interface (2.3), DAC charging control unit (2.4), I/O interface (2.5), PWM respectively.
3. according to the said a kind of multi-functional lead acid accumulator intelligent maintenance device of claim 1, it is characterized in that: said PTC module (3.3) is made up of the parallel connection of some PTC thermistors.
4. a multi-functional capacity of lead acid battery Forecasting Methodology is characterized in that comprising the steps:
(1) foundation of BP neural net
Network configuration is divided into three layers: input layer, hidden layer and output layer, input layer number are 2: discharging current of storage battery
and accumulator voltage
; The implicit number of plies is 1, comprises 6 latent layer unit; Output layer node number is 1: the residual capacity of storage battery
; Hidden node and output node all adopt the Sigmoid function as excitation function, and the training algorithm of network adopts momentum gradient decline back-propagation algorithm;
(2) adopt genetic Annealing algorithm optimization network weight
A) the basic operation parameter of initialization population and network comprises: population scale, the maximum algebraically of evolving, crossover probability, annealing initial temperature
, temperature damping's function and MapkoB chain length;
B) calculate each individual fitness of evaluation;
C) judge whether to satisfy the optimization criterion, then get into step (3) if satisfy, otherwise continue next step;
D) adopt that corresponding genetic operation operator is selected, intersection and mutation operation;
E) optimization of annealing is returned step b) then and is continued circulation;
(3) image data is trained network
Behind completing steps (2); Utilize neural net to have self study character; The discharging current
and the accumulator voltage
of the storage battery that collects are applied in the training of BP neural net, and training process is following:
F) the maximum frequency of training and the minimal error of network are set;
G) with the discharging current
of storage battery and accumulator voltage
training sample as network; Fan-in network adopts momentum gradient decline back-propagation algorithm to train, up to smaller or equal to minimal error or till reaching maximum frequency of training;
6. according to the said a kind of multi-functional capacity of lead acid battery Forecasting Methodology of claim 4, it is characterized in that: d in the said step (2)) said mutation operation adopts the self adaptation aberration rate, and the variation probability is:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011102312786A CN102324582B (en) | 2011-08-12 | 2011-08-12 | Intelligent maintenance device of multifunctional lead-acid battery and capacity prediction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011102312786A CN102324582B (en) | 2011-08-12 | 2011-08-12 | Intelligent maintenance device of multifunctional lead-acid battery and capacity prediction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102324582A true CN102324582A (en) | 2012-01-18 |
CN102324582B CN102324582B (en) | 2013-10-02 |
Family
ID=45452282
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2011102312786A Expired - Fee Related CN102324582B (en) | 2011-08-12 | 2011-08-12 | Intelligent maintenance device of multifunctional lead-acid battery and capacity prediction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102324582B (en) |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102621500A (en) * | 2012-03-30 | 2012-08-01 | 苏州经贸职业技术学院 | Charge and discharge detection device for lithium battery |
CN102916477A (en) * | 2012-08-06 | 2013-02-06 | 关平 | Multifunctional intelligent accumulator discharge device compatible with multiple voltage grades |
CN104375091A (en) * | 2014-11-18 | 2015-02-25 | 柳州市金旭节能科技有限公司 | Monitoring method for electric vehicle power storage battery |
CN104820859A (en) * | 2015-04-27 | 2015-08-05 | 宁夏大学 | Preprocessing and adaptive genetic simulated annealing algorithm-based phase unwrapping method |
CN104953191A (en) * | 2015-05-07 | 2015-09-30 | 国网河南省电力公司电力科学研究院 | Restoration method for performances of transformer substation returned lead-acid storage battery |
CN105870885A (en) * | 2016-05-24 | 2016-08-17 | 北京新能源汽车股份有限公司 | BMS power supply circuit protection device and method and electric automobile |
CN106025407A (en) * | 2016-06-21 | 2016-10-12 | 天能电池集团有限公司 | Method for restoring capacity of electric bicycle batteries |
CN106058335A (en) * | 2016-07-04 | 2016-10-26 | 北京嘉捷恒信能源技术有限责任公司 | Battery discharge device and battery discharge method |
CN106501721A (en) * | 2016-06-03 | 2017-03-15 | 湘潭大学 | A kind of lithium battery SOC estimation method based on biological evolution |
CN108267696A (en) * | 2018-04-09 | 2018-07-10 | 广东电网有限责任公司 | A kind of accumulator capacity shallow discharge detection device |
CN109031152A (en) * | 2018-10-15 | 2018-12-18 | 重庆科技学院 | Lead-acid accumulator residual capacity flexible measurement method based on neural network transparence |
CN109633464A (en) * | 2018-11-26 | 2019-04-16 | 国网浙江省电力有限公司台州供电公司 | Battery sound state discharge test method |
CN110084445A (en) * | 2019-06-04 | 2019-08-02 | 龙源(北京)风电工程技术有限公司 | A kind of Wind turbines battery replacement cycle calculation method for variable pitch |
CN110082682A (en) * | 2019-03-12 | 2019-08-02 | 浙江大学 | A kind of lithium battery charge state estimation method |
CN110622348A (en) * | 2017-05-03 | 2019-12-27 | 株式会社半导体能源研究所 | Neural network, electric power storage system, vehicle, and electronic device |
CN111048859A (en) * | 2014-05-02 | 2020-04-21 | 杜克斯有限公司 | Battery condition determination |
CN111555001A (en) * | 2020-05-20 | 2020-08-18 | 长春光华学院 | Battery cooling system of hybrid electric vehicle and control method thereof |
CN112956064A (en) * | 2018-06-19 | 2021-06-11 | B·E·蔡尔 | Specific-class industrial battery optimization and repair equipment with battery diagnosis and battery life prediction and artificial intelligence device |
CN114925515A (en) * | 2022-05-13 | 2022-08-19 | 武汉理工大学 | New energy electric vehicle-oriented battery health diagnosis method and device and service life prediction method |
CN116759671A (en) * | 2023-08-15 | 2023-09-15 | 深圳市贝视特科技有限公司 | Intelligent maintenance system of lead-acid storage battery |
CN114925515B (en) * | 2022-05-13 | 2024-07-26 | 武汉理工大学 | New energy electric automobile-oriented battery health diagnosis method and device and life prediction method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003197275A (en) * | 2001-12-27 | 2003-07-11 | Panasonic Ev Energy Co Ltd | Estimating method of polarized voltage of secondary battery, estimating method and device of residual capacity of secondary battery, as well as battery pack system |
CN2899132Y (en) * | 2006-05-31 | 2007-05-09 | 薛卫峰 | Capacity maintainer for plumbous-acid accumulator |
CN101067648A (en) * | 2007-06-15 | 2007-11-07 | 华为技术有限公司 | Method and system for deciding electrolyte type storage battery performance |
-
2011
- 2011-08-12 CN CN2011102312786A patent/CN102324582B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2003197275A (en) * | 2001-12-27 | 2003-07-11 | Panasonic Ev Energy Co Ltd | Estimating method of polarized voltage of secondary battery, estimating method and device of residual capacity of secondary battery, as well as battery pack system |
CN2899132Y (en) * | 2006-05-31 | 2007-05-09 | 薛卫峰 | Capacity maintainer for plumbous-acid accumulator |
CN101067648A (en) * | 2007-06-15 | 2007-11-07 | 华为技术有限公司 | Method and system for deciding electrolyte type storage battery performance |
Cited By (32)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102621500A (en) * | 2012-03-30 | 2012-08-01 | 苏州经贸职业技术学院 | Charge and discharge detection device for lithium battery |
CN102916477A (en) * | 2012-08-06 | 2013-02-06 | 关平 | Multifunctional intelligent accumulator discharge device compatible with multiple voltage grades |
CN102916477B (en) * | 2012-08-06 | 2015-02-25 | 关平 | Multifunctional intelligent accumulator discharge device compatible with multiple voltage grades |
CN111048859A (en) * | 2014-05-02 | 2020-04-21 | 杜克斯有限公司 | Battery condition determination |
CN104375091A (en) * | 2014-11-18 | 2015-02-25 | 柳州市金旭节能科技有限公司 | Monitoring method for electric vehicle power storage battery |
CN104820859A (en) * | 2015-04-27 | 2015-08-05 | 宁夏大学 | Preprocessing and adaptive genetic simulated annealing algorithm-based phase unwrapping method |
CN104820859B (en) * | 2015-04-27 | 2017-09-05 | 宁夏大学 | Phase unwrapping method based on pretreatment and Adaptive Genetic simulated annealing |
CN104953191B (en) * | 2015-05-07 | 2017-06-16 | 国网河南省电力公司电力科学研究院 | Fortune performances of the lead-acid battery restored method moves back in a kind of transformer station |
CN104953191A (en) * | 2015-05-07 | 2015-09-30 | 国网河南省电力公司电力科学研究院 | Restoration method for performances of transformer substation returned lead-acid storage battery |
CN105870885A (en) * | 2016-05-24 | 2016-08-17 | 北京新能源汽车股份有限公司 | BMS power supply circuit protection device and method and electric automobile |
CN105870885B (en) * | 2016-05-24 | 2019-05-03 | 北京新能源汽车股份有限公司 | BMS power supply circuit protection device and method and electric automobile |
CN106501721A (en) * | 2016-06-03 | 2017-03-15 | 湘潭大学 | A kind of lithium battery SOC estimation method based on biological evolution |
CN106025407A (en) * | 2016-06-21 | 2016-10-12 | 天能电池集团有限公司 | Method for restoring capacity of electric bicycle batteries |
CN106025407B (en) * | 2016-06-21 | 2019-01-18 | 天能电池集团有限公司 | A kind of method of battery of electric bicycle capacity restoration |
CN106058335A (en) * | 2016-07-04 | 2016-10-26 | 北京嘉捷恒信能源技术有限责任公司 | Battery discharge device and battery discharge method |
US11594770B2 (en) | 2017-05-03 | 2023-02-28 | Semiconductor Energy Laboratory Co., Ltd. | Neural network, power storage system, vehicle, and electronic device |
CN110622348A (en) * | 2017-05-03 | 2019-12-27 | 株式会社半导体能源研究所 | Neural network, electric power storage system, vehicle, and electronic device |
US11955612B2 (en) | 2017-05-03 | 2024-04-09 | Semiconductor Energy Laboratory Co., Ltd. | Neural network, power storage system, vehicle, and electronic device |
CN108267696A (en) * | 2018-04-09 | 2018-07-10 | 广东电网有限责任公司 | A kind of accumulator capacity shallow discharge detection device |
CN112956064A (en) * | 2018-06-19 | 2021-06-11 | B·E·蔡尔 | Specific-class industrial battery optimization and repair equipment with battery diagnosis and battery life prediction and artificial intelligence device |
CN109031152A (en) * | 2018-10-15 | 2018-12-18 | 重庆科技学院 | Lead-acid accumulator residual capacity flexible measurement method based on neural network transparence |
CN109633464A (en) * | 2018-11-26 | 2019-04-16 | 国网浙江省电力有限公司台州供电公司 | Battery sound state discharge test method |
CN110082682B (en) * | 2019-03-12 | 2020-04-24 | 浙江大学 | Lithium battery state of charge estimation method |
CN110082682A (en) * | 2019-03-12 | 2019-08-02 | 浙江大学 | A kind of lithium battery charge state estimation method |
CN110084445A (en) * | 2019-06-04 | 2019-08-02 | 龙源(北京)风电工程技术有限公司 | A kind of Wind turbines battery replacement cycle calculation method for variable pitch |
CN110084445B (en) * | 2019-06-04 | 2020-11-13 | 龙源(北京)风电工程技术有限公司 | Method for calculating replacement period of storage battery for pitch control of wind turbine generator |
CN111555001A (en) * | 2020-05-20 | 2020-08-18 | 长春光华学院 | Battery cooling system of hybrid electric vehicle and control method thereof |
CN111555001B (en) * | 2020-05-20 | 2021-04-02 | 长春光华学院 | Battery cooling system of hybrid electric vehicle and control method thereof |
CN114925515A (en) * | 2022-05-13 | 2022-08-19 | 武汉理工大学 | New energy electric vehicle-oriented battery health diagnosis method and device and service life prediction method |
CN114925515B (en) * | 2022-05-13 | 2024-07-26 | 武汉理工大学 | New energy electric automobile-oriented battery health diagnosis method and device and life prediction method |
CN116759671A (en) * | 2023-08-15 | 2023-09-15 | 深圳市贝视特科技有限公司 | Intelligent maintenance system of lead-acid storage battery |
CN116759671B (en) * | 2023-08-15 | 2024-02-23 | 深圳市贝视特科技有限公司 | Intelligent maintenance system of lead-acid storage battery |
Also Published As
Publication number | Publication date |
---|---|
CN102324582B (en) | 2013-10-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102324582A (en) | Intelligent maintenance device of multifunctional lead-acid battery and capacity prediction method | |
CN104538999B (en) | Photovoltaic generating energy storage system and energy scheduling method thereof | |
CN110525269B (en) | SOC battery pack balance control method | |
CN101976866B (en) | Balanced judgment and supplementary device of energy transfer type battery pack and method thereof | |
CN111366848A (en) | Battery health state prediction method based on PSO-ELM algorithm | |
CN107134827B (en) | Bus type lithium battery group equilibrium system prediction control method | |
Wu et al. | Dynamic battery equalization with energy and time efficiency for electric vehicles | |
CN113176511B (en) | Energy storage charging and discharging optimization method and system considering health state | |
CN105226744A (en) | A kind of power battery pack balance charge/discharge control method based on SOC and system | |
CN103329390B (en) | Chargeable cell system and rechargeable battery system operational | |
CN102255114B (en) | Method and device for uniform charge and discharge of batteries | |
CN102104257A (en) | Energy storage system of apartment building, integrated power management system, and method of controlling the system | |
CN205693405U (en) | Set of cells or capacitor group management system | |
CN107015158B (en) | Lithium battery remaining life on-line quick detection method based on constant-current discharge segment | |
CN106300545A (en) | A kind of active equalization for liquid metal cell controls device and control method | |
CN105119312A (en) | A photovoltaic energy storage scheduling method based on mixed integer non-linear programming | |
CN109004696A (en) | A kind of substation battery multifunction control system and its control method | |
CN107769335A (en) | A kind of multi-mode lithium battery intelligent charging management method and device | |
CN113206508A (en) | Prediction-based active equalization method for energy storage battery of micro-grid | |
CN102751790A (en) | Hybrid energy storage system based on supercapacitor in solar photovoltaic system | |
CN114865668A (en) | Energy storage scheduling support evaluation method | |
CN106786909B (en) | A kind of equalising control device in groups and control method of energy storage liquid metal cell | |
CN103997241A (en) | Multi-level-converter capacitance-voltage-sharing method based on H bridge module | |
Ekanayake et al. | Active and passive based hybrid cell balancing approach to series connected lithium-ion battery pack | |
CN111762059B (en) | Multivariable fusion battery pack balancing method considering battery charging and discharging working conditions |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20131002 Termination date: 20170812 |