Embodiment
The present invention is described in further detail below in conjunction with the accompanying drawings.
First, improved SOC algorithm for estimating
Under existing technical conditions, the inconsistency of lithium battery group mainly produces in process of production.Meanwhile electric automobile
On lithium battery group working environment it is typically all relatively severe, cause inconsistency to be aggravated.This inconsistency would generally pass through electricity
Voltage, capacity and the parameter of internal resistance in pond show, and there are very big shadow in the life-span and performance for battery pack
Ring.Before Balance route is carried out to battery pack, how to carry out evaluation to the coherency state of lithium battery group is a prerequisite bar
Part, the quantitatively evaluating of uniformity provide important data for the equilibrium and maintenance of battery pack and supported.
Electric automobile is can continual mileage come characterize automobile by estimating the state-of-charge SOC of battery pack.SOC is as electricity
The state parameter of pond pool-size, it reflects the residual capacity state of battery pack, and it is numerically defined as battery remaining power
Account for the ratio of battery demarcation total capacity:
In formula one, QRIt is remaining charge capacity in battery, C is nominal (specified) charge capacity of battery.
From the point of view of the above-mentioned SOC definition modes provided, what SOC was represented is between battery remaining power and battery active volume
Ratio, if during using Appreciation gists of the SOC as inconsistency, active volume is deposited between can avoiding each cell
Difference, its target be allow all cells can reach simultaneously discharge and recharge by voltage, ensure to utilize electricity as far as possible
The active volume of pond group.Ensure that the SOC uniformity of all cells is actually equivalent to ensure that all battery discharges are deep
Degree it is consistent, so as to avoid because whole battery pack caused by some cell degree of aging is excessive in battery pack
The problem of even scrapping can be declined.SOC difference can reflect the inconsistency of battery, while SOC estimation can also be
Obtained in real time in battery pack running, so the balance policy based on SOC can suppress the inconsistency of battery pack in real time.
When using SOC estimations, maximum difficult point is exactly the accuracy and real-time for SOC estimations.Put in lithium battery
At the initial stage of electricity, the SOC of battery is varied less, can be because SOC estimates in battery discharge latter stage if it can not accurately estimate
The accumulation of error and cause error excessive.Now if equilibrium again, the challenge for balancing control circuit is quite big
, while balanced effect can also have a greatly reduced quality.In addition, the real-time estimation for SOC is to need certain amount of calculation, work as electricity
When number of batteries is more in the group of pond, to ensure SOC real-time, it is necessary to which equalizing system, which can use, has certain computing capability
MCU.
The dump energy of inside battery direct measurement of having no idea obtains, and is merely able to the number that can be measured by battery
Estimate the dump energy of battery, such as voltage, electric current and internal resistance data according to amount.But, the chemical characteristic right and wrong of inside battery
Often complicated, so linear or other simple function relations will not also be presented between these data and SOC, proposed at present
The method of estimation for SOC always the defects of more or less be present, this also cause SOC state estimations for battery pack into
For the emphasis and difficult point in battery management system.SOC methods of estimation of the prior art are as follows:
(1) ampere-hour method
Ampere-hour method is the most frequently used to estimate the method for battery SOC parameter, the core concept of the method can with formula two come
Statement:
Wherein SOCiIt is the SOC of battery current state;CNIt is the rated capacity of battery;I is battery current;η is imitated for discharge and recharge
Rate.Integration by calculating current to the time, the electricity being lost in a period of time can be obtained.Initial SOC known to cooperation0,
It can be obtained by SOC sometime.The principle of the method and realization are all very simple, but it has the defects of following:First
It is higher to current measurement required precision, if current measurement will definitely not cause SOC calculation errors, long term accumulation, the gradual quilt of error
Amplification;Secondly, for initial SOC0Estimation is also required to the acquisition of certain method;Finally, temperature is higher or current fluctuation
When larger, this method application condition is big.Therefore, in order to obtain reliable SOC estimation, it is necessary to be passed using high performance electric current
Sensor needs enough data to be used to estimate initial state to obtain accurate current value.
(2) open circuit voltage method
The electromotive force of battery may be considered by three parts come what is formed, include the open-circuit voltage (open- of battery
Circuit voltage, OCV), the polarizing voltage of the ohmic voltage of battery and battery.Battery in a non-operative state, and
After prolonged stand, the ohmic voltage and polarizing voltage of battery can all drop to 0, now the open-circuit voltage of battery
The terminal voltage of OCV and battery is equal, is exactly the electromotive force of battery, it is possible to estimates according to the relation curve between OCV-SOC
Calculate SOC.On in real time, there is fairly good linear relationship between OCV-SOC on lead-acid battery, can be with the method come more smart
True estimation SOC.But for lithium battery, linear relationship between the two is just without so obvious, it is therefore desirable to establishes
Play more complicated comparison table.
The significant drawback of open circuit voltage method is needed battery sufficient standing before measuring, and duration needs several hours even ten
Several hours, this causes difficulty to measurement;Secondly, the length of time of repose is also more doubt place;Meanwhile with
The aging of battery, the corresponding relation between open-circuit voltage OCV and SOC can also change.These reasons cause can not in actual use
SOC can be obtained online using open circuit voltage method.Open circuit voltage method can combine ampere-hour method, as SOC value initial in ampere-hour method
Preparation method.Can not possibly be total but the relation between open-circuit voltage and SOC can change with cell degradation, during actual use
It is the relation for remeasuring both this parts, so needing corresponding relation between a kind of dynamic amendment open-circuit voltage OCV and SOC
Algorithm.
(3) load method
The principle of load method and open circuit voltage method are consistent, are to overcome open circuit voltage method to be unable to estimation on line
The shortcomings that battery SOC and propose.Its principle is as follows:If the internal resistance r and operating current I of battery can be obtained, pass through survey
The voltage U at amount load R both ends, the balance electric gesture EMF of battery is calculated according to formula below can.
EMF=U+I*r (formula three)
From analyzing it is recognised that the corresponding relation between EMF and SOC is exactly in open circuit voltage method for open circuit voltage method
Relation between OCV and SOC, so being aware of the SOC that can after EMF correspondingly obtains battery.
For in theory, this method overcomes the shortcomings that open circuit voltage method can not measure SOC in real time really, but actually uses
Middle the method still there is it is obvious the defects of:First, influence that internal resistance of cell r factor is very more, while battery is interior
Resistance inherently has inconsistency, and the internal resistance between each cell might have very big difference, so accurately obtaining
The internal resistance in power taking pond is relatively difficult;Secondly, this method is based on open circuit voltage method, so open circuit voltage method institute in itself
In face of the problem of this method can not avoid.Load method is seldom used to obtain online in electric automobile field
SOC, on the contrary it is often used as the basis for estimation of battery charging and discharging cut-off.
(4) internal resistance method
The basic thought of internal resistance method is consistent with open circuit voltage method, and substantial amounts of experimental study shows the exchange resistance of lithium battery
There is close relationship between anti-or DC internal resistance and the SOC of battery, if it is possible to determine this by some battery samples
The relation of sample, then can obtains the SOC of battery by detecting the internal resistance of battery.
The internal resistance of battery can be divided into AC impedance and DC internal resistance.AC impedance is the resistance for reflecting battery to alternating current
Ability, the value can be measured by alternating current impedance instrument.Similar, DC internal resistance then indicates battery for the anti-of direct current
Anti- ability, the value can then be obtained by detecting in the short time changing value of voltage and the changing value of electric current.
But also it refer in the analysis of load method before, the relation between the internal resistance of battery and SOC is very
Complexity, it be limited by incessantly SOC influence, also influenceed by Multiple factors such as temperature and battery healths, thus internal resistance with
Relation between SOC is difficult really to decide.Meanwhile the internal resistance of battery is often very small, only milliohm rank, this is just right
The required precision of measurement is very high, and the error of measurement is also very big on result influence.So actual answer electric automobile is actual
It can be seldom used to middle this method.
(5) neural network
Neural network is on the premise of network model is built up, and data training is carried out by substantial amounts of sample, so as to come
Estimate battery SOC.Battery is the system of a nonlinearity, and neural network has nonlinear fundamental characteristics, can be compared with
The non-linear kinetic characteristic of good simulation battery, therefore, neural network estimation SOC have preferable effect.And neutral net
Method estimation SOC needs to train substantial amounts of sample data, wherein the data from the sample survey and the method for training trained will all influence its estimation
As a result, another defect is that neural net method needs substantial amounts of resource, and the design for battery management system proposes higher
Requirement, generally require the control chip using higher performance, substantially increase cost.
(6) Kalman filtering method
Kalman filtering method is using battery system as a nonlinear dynamical system, and the wherein SOC of battery is only should
A state in system, accordingly establishes battery model, state equation and observational equation is listed according to model, using extension
Kalman filtering method estimates battery SOC.The basic thought of this method is that the state of dynamical system is made in minimum variance most
Excellent estimation.The problem of which solving inaccurate SOC initial estimates in ampere-hour method and cumulative errors.If it can accurately establish electricity
Pool model, then Kalman filtering method can will accurately estimate the SOC of battery.But there is also following problem for this method:
Estimate that the accuracy of battery SOC depends critically on the accuracy of battery model first, secondly because being applied in Kalman filtering method
Substantial amounts of matrix operation, its speed of service requirement to system processor also just improve.
By summarizing existing SOC estimation method, conventional having is following several:Ampere-hour method, open circuit voltage method, internal resistance method,
Neural network and Kalman filtering method.Any of the above method is under different use environments and for different power electrics
Pond, respectively have each institute it is exclusive the characteristics of, but simultaneously there is also it is different the defects of with deficiency, table 1 is that common SOC is estimated
The advantage and disadvantage of meter method are summarized.
The common SOC methods of estimation of table 1 are summarized
In summary, commonly use needs to estimate SOC under battery off working state in the open circuit voltage method in SOC algorithms, therefore
It is difficult to the requirement for meeting electric powered motor estimation on line SOC, estimation SOC is suitably combined with other method.Ampere-hour method presence pair
The problem of dependence of initial value and increasing cumulative errors and the self discharge problem that battery can not be tackled.Kalman filtering
Method solves the problems, such as in ampere-hour method that SOC initial estimates are inaccurate and cumulative errors, and at the same time, its existing subject matter is
It is very strong to battery model dependence, and the SOC of battery could accurately be estimated by only establishing accurate battery model;
And due to having used substantial amounts of matrix operation in Kalman filtering method, its rate request to system processor is higher.And
Neural network estimation SOC subject matter is due to need to train substantial amounts of sample data, therefore estimation error will be instructed
Practice the influence of data from the sample survey and training method, another defect is that neural net method needs substantial amounts of resource, for cell tube
The design of reason system proposes higher requirement.
In actual use, the phenomenon of aging occurs with the increase of cycle-index in lithium battery.Aging of lithium battery phenomenon
Show the change of inside battery parameter, it is most important that the decay of battery capacity.The lithium electricity of different cycle-indexes can be contrasted
Relation between pond open-circuit voltage and SOC finds out such decay, and 3400mAh ferric phosphate lithium cell is tested in the present invention
In new battery and relation between OCV and SOC after 500 circulations is experienced, the result of experiment is as shown in Figure 1.
From figure 1 it appears that certain corresponding relation is showed between the open-circuit voltage OCV and SOC of lithium battery, but
At the same time the progressively aging of lithium battery can cause the relation between open-circuit voltage OCV and battery SOC gradually to change.With filling
The gradual aging of increase battery of discharge cycles, the capacity of battery constantly reduce.From Fig. 1 it can also be seen that in same open circuit
Voltage OCV, different SOC is corresponded on the OCV-SOC curves of new battery and aged battery.Therefore, make using open-circuit voltage
For SOC basic algorithm for estimating when need to add the change that certain correction algorithm tackles relation between OCV and SOC.
In view of the influence of aging of lithium battery factor, simultaneously because embedded chip leads to used by lithium battery management system
It is weak often to calculate performance comparision, so the improved SOC algorithm for estimating proposed in the present invention has merged ampere-hour method and open-circuit voltage
Method, while certain dynamic corrections algorithm is added, SOC estimated values are maintained in acceptable error range, to tackle
The influence of aging of lithium battery.
As described above, ampere-hour method is initial SOC estimation and use as a kind of conventional method, its error source
During electric current measurement accuracy.The measurement accuracy of electric current is improved by changing high-precision measurement device, and high-precision
Initial SOC can be obtained by open circuit voltage method measurement.But due to the aging effect of battery, the open-circuit voltage and SOC of battery
Between corresponding relation be dynamic change, so the core of improved SOC algorithm for estimating proposed by the present invention be exactly dynamic tie up
Protect SOC and OCV mapping table.During battery use, if in the measured value and mapping table of battery SOC
The difference of value just corrects the mapping table when being more than default error threshold.
The charge and discharge process of battery is divided into three phases in the present invention:Before discharge and recharge, during discharge and recharge and after discharge and recharge.By
Needed in open-circuit voltage measurement by the battery standing long period, so open circuit voltage method is only suitable for battery in non-working condition
In the case of estimate SOC, i.e., the stage before discharge and recharge or after discharge and recharge.The SOC in the two stages is by searching SOC
Obtained with OCV mapping table.In discharge and recharge, SOC can not be directly obtained, but can calculate and release by ampere-hour method
Or the charge value being filled with, then coordinate the SOC before discharge and recharge to obtain.It can be calculated after discharge and recharge terminates by ampere-hour method
SOC after to discharge and recharge.Stand simultaneously after, it can also table look-up to obtain a theoretic SOC by open circuit voltage method.Calculate
Deviation is had between obtained SOC and theoretic SOC, updates SOC's and OCV if difference is more than default error threshold
Mapping table.
The flow chart of improved SOC algorithm for estimating of the present invention is as shown in Figure 2.Specially:
Step 11, judge whether battery is in running order, if it is, step 12 is entered, if it is not, then entering step
Rapid 17;
Step 12, the SOC of current state is calculated using formula twoi, wherein, SOCiIt is the SOC of battery current state, SOC0
For battery functions state when initial SOC, CNIt is the rated capacity of battery, I is battery current, and η is efficiency for charge-discharge,
When charging, η is negative, and when electric discharge, η is positive number;
Step 13, judge whether battery is in running order, if it is, return to step 12, if it is not, then into step
14;
Step 14, open-circuit voltage OCV is measured1, SOC and OCV mapping table is searched, is obtained and the open-circuit voltage
OCV1Corresponding SOC1;
Step 15, SOC is calculatediAnd SOC1Between difference e, when e absolute value is more than default error threshold, into step
Rapid 16, when e absolute value is less than or equal to default error threshold, export SOCi;Into step 11;
Step 16, revised SOC, the output revised SOC amendments are calculated, while it is corresponding with OCV to update SOC
Relation table;Into step 11;
Step 17, open-circuit voltage OCV is measured2, SOC and OCV mapping table is searched, is obtained and the open-circuit voltage
OCV2Corresponding SOC2, export SOC2;Into step 11.
For example, there is battery discharge, its open-circuit voltage can be measured before electric discharge, then by searching SOC and OCV
Mapping table, it is known that SOC now, is designated as S1.When battery starts electric discharge, record current starts to calculate consumption
Electricity, the total electricity of releasing is finally obtained, is designated as Sc.By calculating, it is known that the electricity of last moment, S2=S1-Sc.
When electric discharge terminates, measuring data in contrast table by open-circuit voltage obtains theoretic SOC, is designated as S3.Error amount Sb is defined
Difference between S3 and S2, i.e. Sb=S3-S2.When Sb is beyond default error threshold, then starts makeover process, correct SOC
With the data in OCV mapping table.
When beginning to use algorithm, it is necessary to establish initial a SOC and OCV mapping table before step 11.
In order to obtain such table, while the limitation of space and computing capability is also contemplated for, so have selected interpolation method to obtain this
The table of sample.20 points, i.e., the value of each 5% SOC measurement open-circuit voltages can be selected.Need it is more intensive if can also select
It is even more more to select 50 points, this is depending on demand and chip can use storage.
The mapping table of amendment SOC and OCV in step 16 can be completed by formula four and five.
E=SOCTable-SOCCal(formula four)
Sn(i+1)=Sn(i)-F (n, e) (0≤n≤50) (formula five)
SOC in formula fourTableIn be SOC original in table, and SOCCalIt is then the SOC calculated, has finally used one
Individual e represents both differences.S in formula fiven(i) value in the table after current ith renewal, S are representedn(i+1) it is exactly then more
Value after new in i+1 time table.F (n, e) is exactly correction factor, and the coefficient is a function related to n and e.In the present invention
It is using the expression of linear function:
F (n, e)=a*e*n (formula six)
A therein is an adjustable constant, amendment speed is represented, from figure 1 it appears that in the electricity of SOC difference sections
Press corresponding relation different, so influenced by aging also different under different SOC.Therefore, the amendment speed at different sections should
Differ, i.e. a should take different values at different sections.N is the points of interpolation.If desired, F (n) functional form
It can change, the functional form of renewal can be changed according to different battery behaviors.
To verify the validity of the SOC algorithm for estimating of amendment proposed by the present invention and practicality, it is necessary to verify battery more
Currently whether real corresponding table is consistent with battery by the SOC and OCV mapping table being calculated after secondary discharge and recharge,
Error is in what scope.Therefore, the present invention is devising validity of the experiment porch for verifying the SOC algorithm for estimating,
The schematic diagram of the platform such as Fig. 3.For the present invention with rated capacity 3400mAh, rated voltage is 3.7V 18650 lithium batteries
Monomer is test object.In order to which visible obvious decay, the number of battery charging and discharging circulation need more, this reality in experiment
Circulated in testing 500 times.Simultaneously in order to avoid influence of the temperature to test result, temperature is controlled in test at 25 degrees Celsius.For
Battery can return to static condition really, need battery standing more than 30 minutes when open-circuit voltage is measured.Simultaneously
In order to reduce influence of the pressure drop of internal resistance in electric discharge, battery has used low rates of discharge pattern in electric discharge, so voltage
Curve can be used directly to estimate open-circuit voltage curve.
The result of test is as shown in Figure 4.It can be seen from figure 4 that have between new two curves of battery and aged battery compared with
Big gap.If not adding correction algorithm, the relation between the voltage and capacity of new battery can be still used after cell degradation,
This can cause the error in SOC estimation can be increasing with the continuous aging of battery, be highly detrimental to extend making for battery
Use the life-span.After modification method is added, the curve of calculated curve and aged battery has quite good matching degree, this
Show the aging with battery, corresponding relation between the algorithm dynamic amendment voltage and capacity in the present invention.This can be great
Help to reduce the error that cell degradation is brought to SOC estimations.
2nd, Balance route strategy
As described above, SOC algorithm for estimating is one of core technology of balance policy in balancing technique, it is above-mentioned determining
Amendment SOC algorithm for estimating after, the balance policy that uses of the present invention is introduced below.Balance policy and equalizing circuit
Topology design is most important two key points in balance control system, and equalizing circuit topology needs to coordinate rational balanced plan
The portfolio effect that slightly competence exertion has gone out, both are the complementary two parts of balance control system.Base is employed in the present invention
In the equalizing circuit of two-way flyback transformer, while it is a set of balance policy of the circuit design.
It is more preferable to realize that the present invention extends battery in order to obtain good portfolio effect according to analysis and research above
The design object in group life-span, variables of the SOC as Balance route is employed in of the invention, needs to use in balance control system
Quantitative criteria is turned on or off equalization function.In order to evaluate and identify the inconsistent state of battery pack, the present invention analyzes
Variance δsoc 2With extreme difference rsocTwo different quantizating index, it is used consistent in the two expression formula and statistical mathematics, specifically
It is as follows:
rSOC=max (SOC (i))-min (SOC (i)), i=1...n (formula nine)
Analyzed from statistical mathematics angle, the dispersion degree of SOC value is the side of passing through between each cell in battery pack
Difference is come what is represented, i.e., each cell is smaller relative to average cell SOC dispersion degrees in the smaller explanation battery pack of variance, electricity
The otherness of SOC values is smaller in the group of pond;, whereas if variance is bigger, then illustrate that each cell is relative to average in battery pack
The dispersion degree of SOC value is higher, and the SOC value otherness of each cell is bigger.If so using variance as evaluation lithium electricity
If the quantitative criteria of pond group uniformity, in theory for can obtain good effect.But the meter of variance or standard deviation
Calculation amount is sizable, it is contemplated that it is not powerful often to calculate performance for embedded chip used in balance control system, and comments
The uniformity of valency battery pack needs often or even made in real time again, so evaluated using variance as system conformance
Quantitative criteria is less suitable.
What extreme difference represented is the difference between SOC maximum in battery pack and minimum SOC, when the value is smaller, illustrates electricity
Difference in the group of pond between maximum SOC and minimum SOC is small, that is to say, that the SOC of each cell is distributed in less scope
It is interior, it may be said that the uniformity of bright battery pack is preferable.Gap is very when extreme value is larger, between battery pack maximum SOC and minimum SOC
Greatly, the SOC of battery pack is likely distributed in the wide scope of comparison, illustrates that consistency of battery pack may be poor.Meanwhile calculate pole
The SOC situations of all cells of battery pack need not be taken into consideration when poor, it is only necessary to which finding minimum and maximum SOC value can
With amount of calculation that this is significantly less.
From the above analysis it is recognised that the shape of the uniformity of battery pack can be reflected using SOC variance and extreme difference
Condition, so being all suitable using variance and extreme difference as the evaluation criterion of consistency of battery pack is quantified.But consider simultaneously
The computing capability of balance control system, the calculating of system can be greatly reduced using extreme difference, accelerate the reaction speed of system, so
Extreme difference is more appropriate to the use in the present invention compared to variance.
Mainly the mesh of battery pack balancing is realized in based on SOC balance strategy by reducing the SOC differences between battery
, using the SOC of each single lithium battery as main control object, electricity is reduced by the form to cell discharge and recharge
SOC difference between pond.Balancing procedure used in the present invention is as follows:The SOC of all batteries can be measured when starting balanced,
The one of target SOC as equilibrium of selection.It but can as a rule select averageAs the target SOC of equilibrium, make
The efficiency of equilibrium can be improved as target and give full play to the advantage of charge and discharge balancing by the use of average.The present invention is using SOC
During as Balance route means, there is provided a Balance route band (dSOC) to prevent the fluctuation of equilibrium, employed in the present invention
1% as rejection zone control dSOC.Then SOC is higher thanSingle lithium battery carry out equalization discharge, it is same
To less thanBattery pack carry out charge balancing.The process can be illustrated by flow chart, as shown in Figure 5.Specifically
For:
Step 21, the SOC of each single lithium battery in lithium battery group is obtained, especially by above-mentioned improved SOC algorithm for estimating
Obtain the SOC of each single lithium battery;
Step 22, the extreme difference r of battery pack is calculatedsoc, i.e. calculate SOC maximum in each single lithium battery of battery pack and most
Difference r between small SOCsoc;
Step 23, the extreme difference rsocWith the size of default extreme difference threshold value, as the extreme difference rsocIt is default more than described
During extreme difference threshold value, into step 24, as the extreme difference rsocDuring less than or equal to the default extreme difference threshold value, into step 25;
Step 24, the SOC of all single lithium batteries average is selectedAs the target SOC of equilibrium, SOC is less thanSingle lithium battery carry out charge balancing, SOC is higher thanSingle lithium battery discharge
Weighing apparatus, wherein, dSOC is Balance route band, using 1% as rejection zone control dSOC in the present invention;
Step 25, terminate.
Some special nodes should be noted in the process.When battery is in discharge condition, the electricity of constitution difference
Pond can reach blanking voltage into discharge off in advance, and the good battery of constitution can a remaining part electricity, this just leads
The capacity of battery pack has been caused fully to be used.In order to solve this problem, while also the life-span of each battery is protected to make
Can not be in the state put, the solution method in the present invention is, to will enter electric discharge by battery carry out charging and make
Its SOC and other batteries SOC are consistent, no matter whether the battery is in rejection zone.It so can be whole battery pack
In battery reach SOC=0 state simultaneously, make full use of the capacity of battery pack.
Likewise, when battery is in charged state, the battery of constitution difference into full power state and can reach in advance
Blanking voltage, the capacity for the battery pack for being so equally are wasted, so in order to prevent these batteries from entering cut-off electricity in advance
Pressure, the counter-measure taken in of the invention are when having detected that battery will enter charge cutoff, start equalizing circuit and cause
The SOC of the battery is fluctuated near battery SOC so that final all batteries can reach SOC=1 state simultaneously.
3rd, equalizing circuit
The present invention have selected two-way flyback transformer equalizing circuit.Main cause is that the equalizing circuit electric current is big, balanced
Speed is fast, it is possible to achieve two-way equilibrium.The schematic diagram of typical two-way flyback transformer equalizing circuit is demonstrated in Fig. 6.
The balanced essence of flyback transformer is the mutual conversion by electric energy and magnetic energy, realize energy battery cell it
Between bi-directional.When battery pack section cell is more compared to other energy contents of battery, using flyback transformer as energy
Medium transfer is measured, gives the unnecessary energy transmission of wherein battery to whole battery pack;And when the energy of battery pack batteries is compared
When other batteries are less, equally using flyback transformer as energy transfer medium, the energy input of whole battery pack is given should
Cell, prevent that the energy content of battery is too low and battery is damaged.Such a structure has the balanced way of both direction, tool
Body is as follows.
(1) equilibrium of the cell to battery pack (top is balanced)
After balance control system has detected the SOC of all cells, balanced cell is needed for SOC is higher,
Corresponding change switch of the battery is opened, and disconnects other all switches including primary switch, flyback transformer
There is electric current in secondary windings by the way that now electric energy is stored in the form of magnetic energy in the secondary windings;SOC is dropped in the battery
After desired value, time change switch is disconnected, turns on primary switch, such energy can be transferred to armature winding from secondary windings, and magnetic energy is again
Conversion has passed to whole battery pack for electric energy, has thus controlled SOC highest batteries, while by unnecessary energy
It has been transferred in battery pack in other batteries.
(2) equilibrium of the battery pack to cell (bottom is balanced)
After balance control system has detected the SOC of all cells, balanced cell is needed for SOC is relatively low,
Corresponding primary switch is opened, and disconnects all secondary switch, having electric current in the armature winding of flyback transformer leads to
Cross, now electric energy has been deposited into armature winding in primary side in the form of magnetic energy;After enough electric energy have been filled with, disconnect
Primary switch, and time change switch corresponding to minimum SOC is opened, the secondary windings is turned on, such energy just passes from armature winding
Secondary windings is delivered to, magnetic energy goes back to electric energy and has been filled into the battery, can be by the SOC of this process cell
Go up, battery pack entirety SOC also returns to more consistent numerical value.
Equivalent-circuit model and Fig. 8 as shown in Figure 7 can be waited until by carrying out circuit and Analysis of Magnetic Circuit to flyback transformer
Shown equivalent magnetic circuit figure.
When MOS switch pipe turns on, from the left of Fig. 8, it is unshakable in one's determination in magnetic resistance RMWith the magnetic resistance R of leakage fieldSParallel connection, magnetic are equal
Gesture is N1i1, in RMMagnetic flux phi caused by upperMThrough iron core, after disconnecting metal-oxide-semiconductor, the electric current i that is flowed through in secondary windings2Produce
Magnetic flux phiM.It is visible in Fig. 8 right part, moment and N in conversion1i1=N2i2.If N1:N2=1:In 1, Fig. 7
Equivalent-circuit model is converted by flyback transformer equivalent magnetic circuit modeling, the leakage inductance L of primary and secondary in figureSSize
It is identical.Due to there is air gap in flyback transformer, its iron core has a less inductance value, and the leakage inductance in flyback transformer
LSIt can not ignore.After metal-oxide-semiconductor is opened, i0Flow through primary LSWith LM, now secondary no electric current passes through.When metal-oxide-semiconductor turns off, input
L in side leakage senseSWhole energy and LMPart energy consume in uptake pathway (clamp circuit:In flyback transformer
The voltage stress born with reduction switching tube), and LMIn remaining part energy and secondary LSMiddle whole energy leads to
Secondary is crossed to be exported.
In figure 6, when metal-oxide-semiconductor S is opened, input voltage UiPrimary winding both ends are loaded into, are determined according to Lenz
Rule understands just upper negative induced electromotive force, diode D under now secondary windings generation2It can not turn on, therefore electric current can not be secondary
Level circulates in loop.Now equivalent to one inductance of transformer primary side winding.Assuming that the inductance value of armature winding is Lp, it is secondary
The inductance value of winding is Ls, then the electric current that armature winding is flowed through during metal-oxide-semiconductor conducting is:
In t=tonWhen, primary winding current reaches maximum:
When closing metal-oxide-semiconductor, it is upper just lower negative to understand that the polarity of voltage of secondary windings switchs to according to Lenz's law, now two pole
Pipe D2It is switched on, is stored in magnetic energy in transformer and changes for electric energy, have electric current to flow through in secondary windings, electric current is:
Work as t=toffWhen, the electric current of secondary windings reaches minimum value Ismin.Work as IsminWhen=0, metal-oxide-semiconductor stores during turning on
Magnetic field in energy discharge completely, this process is referred to as the discontinuous operating mode of flyback transformer;Work as Ismin>When 0, metal-oxide-semiconductor
The energy being stored in during conducting in magnetic field does not discharge completely, and this process is referred to as the continuous operation mode of flyback transformer.
The current waveform of above two mode of operation is the same as shown in Figure 9.
Magnetic flux in magnetic core of transformer is required to return to original position when each end cycle, and the principle is referred to as magnetic
It is logical to reset principle, remanent magnetism under continuous operation mode be present, magnetic flux during each end cycle should be kept to recover in theory
To initial value, but because magnetic has an iron loss, for coil windings there is also copper loss, this allows for temperature rise in use,
Cause initial value magnetic flux to shift, can not reset, ultimately result in magnetic flux change and enter nonlinear area, inductance value subtracts
Small, current value increase, magnetic core easily reaches saturation state, transformer cisco unity malfunction, causes the very big dangerous of circuit, together
When continuous operation mode under the volume of flyback transformer be also that religion is big, and volume of transformer is smaller under discontinuous operating mode,
And allow larger primary and secondary electric current.Therefore, flyback transformer of the present invention works in interrupter duty mould
Formula.
Flyback transformer primary side and the 6 groups of battery packs of section monomer ferric phosphate lithium cells series connection employed in the present invention
It is connected, each secondary is connected with each cell, and the rated voltage of each cell is 3.6V, therefore transformer primary side works
Voltage is about 18~24V, and secondary operating voltage range is 4.2~3.0, and the operating efficiency of transformer is designed as 80%, work frequency
Rate is 10KHz.
(1) maximum duty cycle
In the case of in general, the delivery efficiency of transformer increases with the increase of dutycycle, but when dutycycle exceedes
When 50%, circuit can produce concussion.Although this phenomenon can be improved by adding harmonic compensation module in circuit,
But if do not select suitable component, and reasonably arranged, now harmonic compensation module possibly can not in circuit
Play a role, cause the working condition of circuit still can be unstable in the case where dutycycle is more than 50%.So transformer
Maximum duty cycle is normally between 40%~50%, and maximum duty cycle of the present invention finally elects 45% as.Actual use accounts for
Empty ratio will also be obtained by emulating.
(2) turn ratio
The turn ratio N of transformer is the ratio between primary transformer coil number of turn Np and secondary coil number of turn Ns.
The primary and secondary number of turn of transformer can not be directly learnt at the beginning of design.According to shown in formula 13 in the present invention
Transformer reflected voltage directly determines that the no-load voltage ratio N of transformer is:
In above formula, reflected voltage VORRepresent to be formed conversely when there is electric current to flow through on vice-side winding and in primary side winding
Voltage, VoRepresent output voltage, VfIt is the pressure drop of metal-oxide-semiconductor.VORIt can be calculated by following formula.
N=5.2 can be calculated by formula 14 and formula 15, take N=5.
It is demonstrated experimentally that when dutycycle is more than 40%, euqalizing current is excessive, does not meet hardware condition, and when dutycycle is
The electric current of secondary is less than 5A when 20%, is not inconsistent with the target of the present invention, so dutycycle should be when balanced at the top of progress
Between 25% and 35%.Meanwhile when dutycycle is more than 35%, transformer works in continuous mode, thus it is also improper, because
This is that dutycycle should be selected between 20% to 30% in bottom equilibrium.Integrate, select 25% or 30% dutycycle
All it is proper scope for top equilibrium and bottom equilibrium.
4th, balance control system
Present invention employs control chip of the K64 chips of the newest release in grace intelligence Pu as MCU ends, newest powerful
Embedded chip, for SOC algorithm for estimating of the present invention and balance policy expansion provide good performance ensure.While with
The chip is core design voltage measurement module, current measurement module, temperature-measuring module and balance module.It is wherein balanced
Module has used the LTC3300-1 chips for aiming at the design of two-way transformer active equalization circuit, by the advantage of integrated chip,
Balance module has further lifting on circuit complexity and cost control.
Lithium battery group usually needs to combine section monomer lithium ion batteries up to a hundred when in use, it is contemplated that the spirit of battery pack distribution
Activity, whole battery pack needs are divided into many small battery packs and are attached separately in different battery cases, while hold to extend
Amount, subsequently can conveniently extend new battery in battery pack, the present invention using modularization extension design.Overall design frame
Figure is as shown in Figure 10.
Whole system has been divided into upper and lower two-stage in design.Secondary MCU realizes the work(carried out and be monitored to this Battery pack
Can, including:Gather the data such as the voltage of each battery, electric current and temperature in this group;By the data feedback of collection to master control
MCU;Calculate the SOC of each cell;Judge whether to need to carry out equilibrium according to the SOC;When needing to carry out balanced time control
The balance module of this group is made to needing to carry out balanced battery progress equilibrium.Main control MCU is responsible for collecting the number of subordinate MCU feedbacks
According to, while the PC of superior transmits the data, to facilitate the data acquisition and debugging to whole battery pack.Main control MCU may be used also
Corresponding battery pack is forwarded to the order for sending PC.The execution flow of main control MCU is as shown in figure 11, secondary MCU execution
Flow is as shown in figure 12.
Had the following advantages that using such classification and modularized design:
(1) scalability of system is improved.Lithium battery group needs to combine the battery of varying number when in use, such as
Automobile vendor can usually carry out to carry out same vehicle the classification of price, no for the needs of sale according to different course continuation mileages
With course continuation mileage need to integrate the lithium battery of varying number in battery pack, if this in same vehicle using different
Balance control system, the R&D costs of early stage are so not only increased, and the maintenance for different balance control systems also needs
Put into more energy and cost.So use grading design can be with the autgmentability of strengthening system.
(2) real-time of system is improved.MCU in each small battery pack only needs to manage battery in this group i.e.
Can, each subordinate MCU amount of calculation is so greatly reduced, enhances the real-time of balance control system.
(3) compatibility of system is improved.If the baby battery group using different vendor is needed in same big battery group, only
Need that the balance module of the group is debugged or designed again, avoid the wholesale revision to whole system.
(4) reliability of balance control system is increased.Modular design can avoid the paralysis of whole system, when some
When balance module in baby battery group breaks down, other baby battery groups remain to normal work.Other small electricity were so both protected
Battery in the group of pond, while be also avoided that and more serious accident occurs.
In terms of specific hardware circuit, the present invention uses following design method:
Power circuit
The effect of power module is that the power supply of normal work is provided for secondary MCU, the power module that the present invention designs
Circuit theory diagrams are as shown in figure 13.Secondary MCU main control chip K64 normal working voltage is led between 1.71V to 3.6V
Ensure that chip power supply also needs to use 5V voltage in 3.3V or so, equalizing system of the invention when often using.Due to secondary
MCU and balance module are all disposed within baby battery group, it is possible to directly from baby battery group power taking.One baby battery group is usual
It is made up of 6 cells, the operating voltage of each battery is between 3.0V to 4.2V, therefore the terminal voltage of baby battery group exists
Between 18V to 25.2V.The system employs the LM2576 voltage conversion chips of NI companies.LM2576 chips can receive 7 and arrive
40V control source, output voltage 5V, 3A load can be driven, Line and load regulation ability is very powerful, simultaneously
Frequency compensator and a fixed oscillator are also integrated with inside LM2576, by seldom external component can
Complete good voltage output.The circuit that a 5V turns 3.3V is further comprises in fig. 13, because K64 usual operating voltage
It is that 3.3V is output to K64 power supplies so needing to use an ASM1117-3.3V power module of voltage regulation that 5V is changed into 3.3V.Circuit
Two power supply indicators of middle addition be used to indicating two kinds of power supplys whether normal work.
Voltage collection circuit
Accurate voltage collection circuit is not only related to normal use and the monitoring of battery pack, and battery is carried out accurate
Equilibrium judge necessary guarantee.So lithium battery is together in series to form battery pack in use, needing to each in battery pack
The cell voltage that economizes on electricity carries out accurate voltage measurement.
Being usually used in the battery voltage acquisition mode of serial lithium battery group includes common mode mensuration and differential mode mensuration.It is wherein common
Mould measurement is relative to same datum, and the mode to be decayed with precision resistance equal proportion measures the voltage of battery pack, Ran Houyi
It is secondary subtract each other after obtain each section monomer battery voltage.The advantages of this method is that circuit is simple, but the measurement accuracy of this method has
Rely and determined in divider resistance, be easily affected by temperature and produce serious cumulative errors, be suitable only for series-connected cell in this way
Number is less and to the less demanding occasion of measurement accuracy.It is unsuitable for the occasion relied on the accurate SOC of voltage calculating.
The method for having selected differential mode to measure in the present invention, this method gate each batteries by certain method and entered successively
Row measurement.This method is suitable for that series-connected cell number is more, while the occasion higher to required precision.Measured using this differential mode
Mode be also an advantage that it is the normal work that other passages can not be influenceed when certain collection all the way is out of order.This
Outside, it is compared to for the acquisition mode using integrated chip, the metering system of this subchannel only needs when breaking down
Corresponding faulty channel is repaired, and is unlikely to change whole chip, is of great advantage for the maintenance cost for reducing the later stage.
(1) cell selection circuit
The mode of differential mode measurement needs that each cell can be gated, and MOSFET PS7241-2A works are employed in the present invention
For the cell gating switch of voltage acquisition.PS7241 family devices are by light emitting diode (input side) and normally opened contact
Metal-oxide-semiconductor (outlet side) forms.Each PS7241-2A includes two separate gating switches, is work the characteristics of the device
Make that electric current is low, pressure voltage is high to have very fast reaction speed simultaneously.The schematic diagram of cell gating circuit switch such as Figure 14
It is shown.
Resistance R1~R4 is current limliting divider resistance in Figure 14, for limiting the size of electric current in circuit in measurement process.When
The voltage for needing to measure some cell is, it is only necessary to by K64 by 1,3 two pin level on corresponding PS7241
Draw high, now the both end voltage of respective battery monomer can export from PS7241 6,8 two pins.With the circuit in Figure 14
Exemplified by, it is necessary to when measuring bat1 voltage, 1 in PS1,3 pins are drawn high, what is now exported from 6,8 is exactly bat1 both ends
Voltage, wherein 6 pin export anode, 8 pin are GND.And when needing to measure battery bat2, by 3 pin in PS1
With the 1 pin gating in PS2, from the output of PS1 6 pin and PS2 8 pin be exactly bat2 both ends voltage, wherein PS2 8 pin are
Battery bat2 positive pole, PS1 6 pin are battery bat1 negative poles.
(2) voltage reversing circuit
In superincumbent analysis it has already been indicated that during measurement bat1 and bat2, for corresponding to CAP_1 and CAP_2 on pin
Voltage be opposite, in fact all odd number batteries and even number the battery voltage in measurement are all opposite.Therefore
It is of the invention to have used two pieces of PS7241 more, devise the voltage reversing circuit such as Figure 15.Measurement odd number can be made by this circuit
Number battery and even number battery are output identical voltage directions, facilitate magnitude of voltage of the follow-up AD circuits to each cell
Measurement.The reason for not using relay is, because the no PS7241 of the reaction of relay is fast, while pressure drop caused by relay
Also can have an impact to the levels of precision of result.
When testing the battery of odd number, for battery just, CAP1 is born CAP2 for battery, now needs to control K_AD_1 to be set to
Low-voltage, K_AD_2 are high voltage, that is, cause the 8 pin conducting of PS5 and PS6 in figure, and 6 are not turned on.When test even number electricity
For battery just, CAP2 bears for battery by Chi Shi, CAP1, now controls K_AD_1 to put high level, K_AD_2 sets low level, now PS5
Turned on PS6 6 pin, and 8 are not turned on.By the circuit can AD_P be connected to anode all the time, and AD_N connects all the time
To the negative pole of battery.
(3) voltage-frequency converting circuit
Voltage-frequency converting circuit (VFC) can be converted to the voltage signal of input in frequency signal output, VFC circuits
The frequency signal and input voltage signal of output are linearly proportional, that is, voltage gets over the frequency of height output also just
It is faster.VFC circuits are widely used in various circuits, including signal frequency modulation, phase modulation, A/D convertor circuit etc..VFC circuits have
There is the advantages of strong antijamming capability, isolation convenience, stable performance, small high sensitivity and nonlinearity erron.Meanwhile to simulation
When signal is digitized processing, the resolution ratio and precision of VFC circuits are all to be higher than the circuit of AD conversion, and in phase
As a rule cost can be lower for VFC circuits under the premise of same precision.
The voltage-frequency converting circuit designed in the present invention has two parts composition.Part I is operational amplification circuit, this
Part core is one piece of OP07C operational amplifier.OP07C has the characteristics of low noise, non-chopper-zero-stabilized.For most
For usage scenario, OP07C does not need outside prototype part to offset zeroing and frequency calibration.In addition, OP07C is also with low inclined
Electric current is put, the characteristics of high open-loop gain and operating temperature range broadness.In order to ensure the stability of VFC circuits and anti-interference energy
The multiplication factor of operational amplifier has been set to 2 times by power, the present invention.
The Part II of voltage-frequency converting circuit is exactly VFC circuits, and the core of the part is AD7740 chips.
AD7740 is one piece of low cost, the minimum voltage to frequency convert chip of volume.This chip can be operated in 3.0V to 3.6V or
Between 4.75V to 5.25V voltage, the minimum reachable 0.9mA of operating current.AD7740 supports the operating temperature range of non-constant width,
The outside prototype part relied on is seldom, and voltage conversion frequency is accurate.Chip internal is integrated with 2.5V reference, also supports
Using outside input VDD as the reference voltage.Chip also has a synchronised clock input pin --- and CLKIN, highest can be with
Support 1MHz frequency input.Using synchronised clock of the K64 output clock as AD7740 in the present invention, so reduce
Unnecessary prototype part, reduce complexity in circuits.
When analog voltage changes between 0V to VREF, AD7740 signal output frequency is at 0.1 to 0.9 times
Interior linear change between FCLKIN.Its voltage and the conversion formula of frequency are as follows:
It is as shown in figure 16 finally to provide voltage-frequency converting circuit.
It is worth noting that, OP07C needs generating positive and negative voltage to power, realized so also needing to an anti-polarity circuit by+5V
Voltage conversion into -5V voltage, for give OP07C provide negative supply.MAX660 charge pump reversed polarity is employed herein
Switch integrated regulator realizes this function, and circuit diagram is as shown in figure 17.
Current collection circuit
Accurate current measurement is using essential condition during ampere-hour method progress SOC estimations, is employed herein
Hall current sensor carries out the measurement of electric current.The principle of Hall current sensor is when primary current flows through a long wire
When, magnetic field is produced around wire, the size in magnetic field is directly proportional to the size of electric current, and caused magnetic field is gathered in magnet ring, leads to
Cross Hall element in magnet ring air gap to measure and amplify output, its output voltage can reflect the size of primary current.
The advantages of Hall current sensor is that measurement range is wide, can measure the electric current and voltage of random waveform, or even right
Peak point current, voltage signal in transient state can be reflected verily.The corresponding speed of Hall current sensor is exceedingly fast,
The reaction speed of us levels can be reached.The precision of Hall current sensor is very high simultaneously, can accomplish the measurement better than 1%
Precision, measurement thread degree, and acceptable long-time unfailing service, can generally ensure up to a few hours
Continuous work.In addition, Hall element can accomplish that the volume use of very little is also convenient for.
Hall current sensor the measurement range reachable ± 100A, operating voltage 5V that the present invention selects.When electric in circuit
It zero is voltage output 2.5V that stream, which is,;When electric current is -100A in single circuit, voltage output 0V;When electric current is in circuit
Output is 5V during 100A.For the sample circuit using OP07C, the output for the Hall current sensor that the present invention uses
In normal range of operation.
Temperature collection circuit
Temperature collection circuit is used to gathering the temperature value of each baby battery group, be during operation of the temperature for lithium battery have compared with
On the one hand big influence, real-time temperature value can ensure the accuracy of the estimation to lithium battery SOC, on the other hand can mistake
High temperature is for the safe for operation and essential of system.
Present invention employs the temperature collection circuit of the designs of the thermistor NTC10KB3950K based on NTC to realize to small electricity
The characteristics of temperature acquisition of pond group, the circuit have that measurement accuracy is high, and Simultaneous Stabilization simple in construction is good.NTC10KB3950K
Precision can be to 1%, and at 0 degree, resistance is 32.5K, and corresponding voltage is 0.29V, and 85 degree are resistance 1.063K, and corresponding voltage is
3.26V.Calculating between resistance and voltage is given by the following formula:
Then the voltage and vs. temperature table according to NTC10K-3950, can check in the temperature value in baby battery group.
Equalizing circuit
Balancing control circuit is one of core in the present invention.The present invention devises a kind of based on LTC3300-1 chips
Transformer equalizing circuit.LTC3300-1 is a controller IC with error protection, suitable for the electricity of multiple batteries composition
Pond group carries out the two-way active equalization based on transformer.The device has been integrated with institute's gate driving circuit in need, high accuracy
The house dog of battery sense, failure detector circuit and a built-in timer.Each LTC3300-1 can utilize a 36V
Common mode input carries out balanced to the serial lithium battery of up to 6 sections.Can be by the electric charge of arbitrarily selected battery with high efficiency
Mode carry out transmission back between the even more adjacent cells of 12 sections at its own.The SPI interface that LTC3300-1 is carried, can
It is electric to long string concatenation so as to realize to complete to connect with multiple LTC3300-1 devices in the case where not using light-coupled isolation
Charge balance in pond per batteries.The LTC3300-1 of series connection can independent operation simultaneously, therefore allow in battery pack
All batteries concurrently and independently carry out balanced management.
Independent operating during equalizer corresponding to each LTC3300-1.The primary side of transformer by a metal-oxide-semiconductor with
Each cell is connected, and the secondary of transformer is then connected by a MOS with whole battery pack.LTC3300-1 supports two kinds
The balanced way of transformer.A kind of is the transformer primary side and secondary that each transformer has oneself;Another kind is all transformations
Device has the primary side of oneself, but shares a secondary being connected with battery pack.The connection signal of the upper two kinds of transformers of LTC3300-1
Figure is as shown in Figure 18;
In Figure 18 (a), the balanced transformer for each cell has independent primary and secondary side, and primary side passes through
One metal-oxide-semiconductor is linked to cell, and secondary is connected to battery pack by a metal-oxide-semiconductor;In Figure 18 (b), each balanced transformation
Device only has single primary side, and by metal-oxide-semiconductor connecting single batteries, secondary is that all transformers share.In view of just individual system
Module design object, and after the angle more easily safeguarded, present invention employs form all in Figure 18 (a).
LTC3300-1 can at most be accessed most can carry out equilibrium to 6 batteries, and Figure 19 gives passage 2 therein
Connection, the solution on other roads are similar with the road.C2 pins are linked to bat2 positive pole in figure, and I2P, I2S are respectively used to survey
The electric current of quantitative change depressor primary and secondary side, G2P and G2S are then used for the metal-oxide-semiconductor break-make for controlling former secondary.BAT+, BAT- are linked to
Be whole battery pack both positive and negative polarity.
It is appreciated that embodiment as described herein can be by hardware, software, firmware, middleware, microcode or its any combination
To realize.For hardware implementation mode, processing unit can be in one or more application specific integrated circuits (ASIC), data signal
Processor (DSP), digital signal processing device (DSPD), PLD (PLD), field programmable gate array
(FPGA), processor, controller, microcontroller, microprocessor, be designed to perform other electronics lists of function described herein
Realized in member or its combination.When realizing embodiment with software, firmware, middleware or microcode, program code or code segment
When, it can be stored in the machine readable media of such as storage assembly.
It is obvious to a person skilled in the art that the invention is not restricted to the details of above-mentioned one exemplary embodiment, Er Qie
In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, nothing
By from the point of view of which point, embodiment all should be regarded as exemplary, and be nonrestrictive, the scope of the present invention is by institute
Attached claim rather than described above limit, it is intended that will fall in the implication and scope of the equivalency of claim
All changes include in the present invention.Any reference in claim should not be considered as to the involved right of limitation
It is required that.
Moreover, it will be appreciated that although the present specification is described in terms of embodiments, not each embodiment is only wrapped
Containing an independent technical scheme, this narrating mode of specification is only that those skilled in the art should for clarity
Using specification as an entirety, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art
It is appreciated that other embodiment.