CN102110859A - Method for restoring lead-acid battery based on time series tracking and forecasting - Google Patents

Abstract

The invention relates to a method for restoring a lead-acid battery based on time series tracking and forecasting, which belongs to the technical field of equipment for restoring the lead-acid batteries. The method comprises the following steps: continuously sampling all single-lattice temperatures Ti(0)(k) of the lead-acid battery during the restoring process in a fixed cycle; computing the temperature rise changes DTi(0)(k) of polar plates between the adjacent single lattices of the lead-acid battery, establishing a p-order autoregressive model AR (p) of the temperature rise changes DTi(0)(k) of the polar plates between the adjacent single lattices of the battery, optimizing the model order of the p-order autoregressive model AR (p) of the temperature rise changes DTi(0)(k) of the polar plates between the adjacent single lattices of the lead-acid battery, tracking and forecasting the temperature rise changes DTi(0)(k) of the polar plates between the adjacent single lattices of the lead-acid battery in advanced based on the p0-order autoregressive model AR (p0), determining whether the forecasted values of temperature rise DTi(0)(k) of all the single lattice polar plates of the lead-acid battery are consistent or not, if so, stopping the restoring process, otherwise, returning to the step 1.1. The restoring process adopts the strategy of performing the time series tracking on the temperature rise between the signal lattices of the lead-acid battery, the restoring degree is determined in a self-adaptive manner, the restoring process is automatically finished by judging the consistency of the temperature rise of all the single lattices, and the problem of over-restoration of the single lattices can be inhibited.

Description

Lead acid accumulator restorative procedure based on the time series tracking prediction

Technical field

The present invention relates to a kind of lead acid accumulator restorative procedure based on the time series tracking prediction.Belong to lead acid accumulator prosthetic appliance technical field.

Background technology

Along with popularizing that lead acid accumulator uses, become the focus problem useful life of lead acid accumulator.And the sulfation of the greatest factor that the causes useful life decay accumulator negative plate that to be exactly long-term undercharge cause; Because the charging effect of all kinds charger is uneven, the sulfation problem of storage battery just inevitably takes place, and the appearance of storage battery prosthetic device has just become the problem that when the water comes, a channel is formed.From the market status, storage battery vulcanization prosthetic device and reparation are theoretical uneven, comparatively popular is that impulse train is repaired in conjunction with the chemistry reparation, but this kind method is long repair time, and uncertain factor is many, it is particularly outstanding to cross the reparation problem, with its serve as main research and development various prosthetic devices repairing effect also difficulty make consumer satisfaction.8 tunnel pulse repair layers that must health power station DK-2008 are example, its pulse frequency group working range is 0.1KHZ～9.9KHZ, pulse current 0.15A～1.6A, temperature limit fixes on 65 ℃ ± 2%, general from its repairing effect of practical application effect, find after tested to exist single lattice to cross the reparation phenomenon.

In general, there is two large problems in its repair process:

1, repair process adopts Passive Control, and the deadline is repaired in promptly artificial decision or timing decision;

2, adopt the whole strategy of repairing in the repair process, occur certain single lattice easily and cross the reparation problem.

Summary of the invention

The technical problem to be solved in the present invention is: the defective at more present lead acid accumulator prosthetic devices provides a kind of lead acid accumulator restorative procedure based on the time series tracking prediction.

The technical solution adopted for the present invention to solve the technical problems is: should be based on the lead acid accumulator restorative procedure of time series tracking prediction, and it is characterized in that: step is as follows

1.1 each single Ge WenduT of lead acid accumulator in the fixed cycle continuous sampling repair process _i ⁽⁰⁾(k);

Wherein: i represents the single lattice of lead acid accumulator i, and i ∈ [1,6], k are sampling instant;

Change DT 1.2 calculate the temperature rise of pole plate between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k), embodying formula is:

${\mathrm{DT}}_i^{\left(0\right)}\left(k\right)=T_i^{\left(0\right)}\left(k\right)-{T_{i-1}}^{\left(0\right)}\left(k\right),i\≥2$

Change DT 1.3 set up the temperature rise of pole plate between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k) p rank autoregression model AR (p), the formula of embodying is:

ε wherein _kBe the white Gaussian noise sequence, p value maximum is got number of samples M one time, Be the model following coefficient;

1.4 the temperature rise of pole plate changes DT between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k) red pond criterion AIC is followed in the model order optimization of p rank autoregression model AR (p), optimization method, and the formula of embodying is:

AIC(p)＝ln(σ ²(p))+2p/M

Wherein, σ ²(p) be p rank model of fit residual error variances;

If there is p ₀∈ [1, p] makes P then ₀Be best autoregression model exponent number;

1.5 the temperature rise based on pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) p ₀Rank autoregression model AR (p ₀), the temperature rise of carrying out pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) tracking prediction in advance, the formula of embodying is:

1.6 determine each single lattice pole plate temperature rise DT of lead acid accumulator _i ⁽⁰⁾Whether predicted value (k) is consistent, promptly

Whether equate,, otherwise get back to step 1.1 if equate then to stop repair process.

Compared with prior art, this based on the beneficial effect that the lead acid accumulator restorative procedure of time series tracking prediction is had is:

Because the non-equilibrium that discharges and recharges has caused the difference of each single lattice state of cure (vulcanization) of lead acid accumulator, in the lead acid accumulator repair process, significantly embodying is exactly the difference that the temperature rise between each single lattice pole plate of lead acid accumulator changes, the present invention is directed to the defective of some lead acid accumulator prosthetic devices existence at present, solved following problem:

1, repair process has embodied the factor of ACTIVE CONTROL, utilizes the time series tracking prediction, obtains in advance to repair degree information, has offset the reparation problem of crossing that system responses lags behind and causes;

2, repair process adopts time series to follow the tracks of the strategy of each single compartment temperature rise of lead acid accumulator, self adaptation is judged the reparation degree, and by the conforming judgement of each single lattice temperature rise, finishes repair process automatically, suppress single lattice and crossed the appearance of reparation problem, solved the problem of Passive Control repair time.

Description of drawings

Fig. 1 is the lead acid accumulator restorative procedure flow chart based on the time series tracking prediction of the present invention;

Fig. 2 is the lead acid accumulator restorative procedure repair process whole structure figure that the present invention is based on the time series tracking prediction;

Fig. 3 is the placement location schematic diagram of temperature sensor of the present invention in storage battery.

1 battery cell case, 2 anticorrosion high-precision sensor electric source line interface 3-8 are anticorrosion high-precision sensor among Fig. 3.

Embodiment

Below in conjunction with accompanying drawing 1-3 the lead acid accumulator restorative procedure that the present invention is based on the time series tracking prediction is described in further detail:

As shown in Figure 1:

Lead acid accumulator restorative procedure based on the time series tracking prediction of the present invention, concrete steps are as follows:

Step 1, each single Ge WenduT of lead acid accumulator in the fixed cycle continuous sampling repair process _i ⁽⁰⁾(k);

Wherein: i represents the single lattice of lead acid accumulator i, and i ∈ [1,6], k are sampling instant.Fixed cycle is 5min～20min, and the sampling sequence element number is no less than four data units.The formula of embodying is:

$T_1^{\left(0\right)}\left(k\right)=\left\{\begin{array}{cccc}{T}_1^{\left(0\right)}\left(1\right)& T_1^{\left(0\right)}\left(2\right)& \·\·\·& \end{array}{T}_1^{\left(0\right)}\left(M\right)\right\}$

$T_2^{\left(0\right)}\left(k\right)=\left\{\begin{array}{cccc}{T}_2^{\left(0\right)}\left(1\right)& T_2^{\left(0\right)}\left(2\right)& \·\·\·& \end{array}{T}_2^{\left(0\right)}\left(M\right)\right\}$

$T_3^{\left(0\right)}\left(k\right)=\left\{\begin{array}{cccc}{T}_3^{\left(0\right)}\left(1\right)& T_3^{\left(0\right)}\left(2\right)& \·\·\·& \end{array}{T}_3^{\left(0\right)}\left(M\right)\right\}$

$T_4^{\left(0\right)}\left(k\right)=\left\{\begin{array}{cccc}{T}_4^{\left(0\right)}\left(1\right)& T_4^{\left(0\right)}\left(2\right)& \·\·\·& \end{array}{T}_4^{\left(0\right)}\left(M\right)\right\}$

$T_5^{\left(0\right)}\left(k\right)=\left\{\begin{array}{cccc}{T}_5^{\left(0\right)}\left(1\right)& T_5^{\left(0\right)}\left(2\right)& \·\·\·& \end{array}{T}_5^{\left(0\right)}\left(M\right)\right\}$

$T_6^{\left(0\right)}\left(k\right)=\left\{\begin{array}{cccc}{T}_6^{\left(0\right)}\left(1\right)& T_6^{\left(0\right)}\left(2\right)& \·\·\·& \end{array}{T}_6^{\left(0\right)}\left(M\right)\right\}$

Wherein M is number of samples in phase weekly.

Step 2, the temperature rise of calculating pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k), embodying formula is:

${\mathrm{DT}}_2^{\left(0\right)}\left(k\right)=T_2^{\left(0\right)}\left(k\right)-T_1^{\left(0\right)}\left(k\right)$

${\mathrm{DT}}_3^{\left(0\right)}\left(k\right)=T_3^{\left(0\right)}\left(k\right)-T_2^{\left(0\right)}\left(k\right)$

${\mathrm{DT}}_4^{\left(0\right)}\left(k\right)=T_4^{\left(0\right)}\left(k\right)-T_3^{\left(0\right)}\left(k\right);$

${\mathrm{DT}}_5^{\left(0\right)}\left(k\right)=T_5^{\left(0\right)}\left(k\right)-T_4^{\left(0\right)}\left(k\right)$

${\mathrm{DT}}_6^{\left(0\right)}\left(k\right)=T_6^{\left(0\right)}\left(k\right)-T_5^{\left(0\right)}\left(k\right)$

Step 3, the temperature rise of setting up pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) p rank autoregression model AR (p), the formula of embodying is:

${\mathrm{\Φ}}^2\left(B\right){\mathrm{DT}}_2^{\left(0\right)}\left(k\right)={\mathrm{\ϵ}}_k$

${\mathrm{\Φ}}^3\left(B\right){\mathrm{DT}}_3^{\left(0\right)}\left(k\right)={\mathrm{\ϵ}}_k$

${\mathrm{\Φ}}^4\left(B\right){\mathrm{DT}}_4^{\left(0\right)}\left(k\right)={\mathrm{\ϵ}}_k$

${\mathrm{\Φ}}^5\left(B\right){\mathrm{DT}}_5^{\left(0\right)}\left(k\right)={\mathrm{\ϵ}}_k$

${\mathrm{\Φ}}^6\left(B\right){\mathrm{DT}}_6^{\left(0\right)}\left(k\right)={\mathrm{\ϵ}}_k$

Wherein,

Be the white Gaussian noise sequence, p value maximum is got number of samples one time

Be model following coefficient, i=2,3,4,5,6.Carry out model optimization respectively at lead acid accumulator every single lattice p rank autoregression model AR (p), the initial value of p generally is taken as and is not more than

Max-int.The model following coefficient

Definite method be least-squares estimation based on following system of linear equations, expression formula is:

Then have

Wherein

$Z=\left(\begin{array}{cccc}{\mathrm{DT}}_i^{\left(0\right)}\left(p\right)& {\mathrm{DT}}_i^{\left(0\right)}(p-1)& \·\·\·& {\mathrm{DT}}_i^{\left(0\right)}\left(1\right)\\ {\mathrm{DT}}_i^{\left(0\right)}(p+1)& {\mathrm{DT}}_i^{\left(0\right)}\left(p\right)& \·\·\·& {\mathrm{DT}}_i^{\left(0\right)}\left(2\right)\\ \·& \·& & \·\\ \·& \·& \·\·\·& \·\\ \·& \·& & \·\\ {\mathrm{DT}}_i^{\left(0\right)}(M-1)& {\mathrm{DT}}_i^{\left(0\right)}(M-2)& \·\·\·& {\mathrm{DT}}_i^{\left(0\right)}(M-p)\end{array}\right)$

Step 4, the temperature rise of pole plate changes DT between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k) red pond criterion AIC is followed in the model order optimization of p rank autoregression model AR (p), optimization method, and the formula of embodying is:

AIC(p)＝ln(σ ²(p))+2p/M

Wherein, σ ²(p) be p rank model of fit residual error variances.

If there is p ₀∈ [1, p] makes

P then ₀Be best autoregression model exponent number.

Step 5 is based on the temperature rise variation DT of pole plate between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k) p ₀Rank autoregression model AR (p ₀), the temperature rise of carrying out pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) tracking prediction in advance, the formula of embodying is:

Step 6 is determined each single lattice pole plate temperature rise DT of lead acid accumulator _i ⁽⁰⁾Whether predicted value (k) is consistent, promptly

Whether equate,, otherwise get back to step 1.1 if equate then to stop repair process.

So-called equate to be meant equate that promptly each sequence prediction value difference in twos is not more than the minimum value of specification error in the microprocessing unit, thinks that then each sequence prediction value is equal to each other in certain error range.

Embodiment 1:

Prosthetic appliance of the present invention is on the basis of 8 tunnel pulse repair layers of DK-2008 (getting health), the auto-power-off device of external microprocessing unit control; That the reparation object is selected for use is the prosperous and powerful non-maintaining Moped Scooter storage battery FC12-12 that Nanchang City, Jiangxi Province powerful power supply Science and Technology Ltd. produces, capacity under 20 hours discharge rates of this storage battery is 12AH, variations in temperature conformity error set point is 0.01, and the placement location schematic diagram of deciding the anticorrosion temperature sensors of high precision of storage battery as shown in Figure 3.

The present invention is based on the elaborating of lead acid accumulator restorative procedure of time series tracking prediction below in conjunction with the present invention is directed to a certain stage of repair process:

The first step, unit period number of samples are 20, each single Ge WenduT of lead acid accumulator in the continuous sampling repair process _i ⁽⁰⁾(k) original series is as shown in table 1:

In second step, the temperature rise of calculating pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k), each single lattice sequence numerical value is as shown in table 2:

In the 3rd step, determine that tentatively the temperature rise of pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) rank of p rank autoregression model AR (p) Round left p=4, try to achieve according to the principle of least square

As shown in table 3:

Wherein, $Z=\left(\begin{array}{cccc}{\mathrm{DT}}_i^{\left(0\right)}\left(4\right)& {\mathrm{DT}}_i^{\left(0\right)}\left(3\right)& \·\·\·& {\mathrm{DT}}_i^{\left(0\right)}\left(1\right)\\ {\mathrm{DT}}_i^{\left(0\right)}\left(5\right)& {\mathrm{DT}}_i^{\left(0\right)}\left(4\right)& \·\·\·& {\mathrm{DT}}_i^{\left(0\right)}\left(2\right)\\ \·& \·& & \·\\ \·& \·& \·\·\·& \·\\ \·& \·& & \·\\ {\mathrm{DT}}_i^{\left(0\right)}\left(19\right)& {\mathrm{DT}}_i^{\left(0\right)}\left(18\right)& \·\·\·& {\mathrm{DT}}_i^{\left(0\right)}\left(16\right)\end{array}\right),$ $D=(\left(\begin{array}{c}{\mathrm{DT}}_i^{\left(0\right)}\left(5\right)\\ {\mathrm{DT}}_i^{\left(0\right)}\left(6\right)\\ \·\\ \·\\ \·\\ {\mathrm{DT}}_i^{\left(0\right)}\left(20\right)\end{array}\right)-\left(\begin{array}{c}{\mathrm{\ϵ}}_5\\ {\mathrm{\ϵ}}_6\\ \·\\ \·\\ \·\\ {\mathrm{\ϵ}}_{20}\end{array}\right));$

The 4th step is according to red pond criterion and formula AIC (p)=ln (σ ²(p))+and 2p/M, definite DT ₂ ⁽⁰⁾(k), ZDT ₃ ⁽⁰⁾(k), ZDT ₄ ⁽⁰⁾(k), ZDT ₅ ⁽⁰⁾(k) and ZDT ₆ ⁽⁰⁾(k) optimal models exponent number, red pond formula value is as shown in table 4, as can be seen from Table 4 according to expression formula

$\mathrm{AIC}\left(p_0\right)=\underset{1\&le;p\&le;M}{\mathrm{<figure-callout\; id="1"\; label="min"\; filenames="H2009102310618E00011.png,H2009102310618E00021.png"\; state="\{\{state\}\}">min</figure-callout>}}\mathrm{AIC}\left(p\right)$

Can judge DT ₂ ⁽⁰⁾(k), ZDT ₃ ⁽⁰⁾(k), ZDT ₄ ⁽⁰⁾(k), ZDT ₅ ⁽⁰⁾(k) and ZDT ₆ ⁽⁰⁾(k) optimal models exponent number is 3.

The 5th step is based on the temperature rise variation DT of pole plate between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k) 3 rank autoregression model AR (3), the temperature rise of carrying out pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) tracking prediction in advance, the model following coefficient is as shown in table 4, and its predictive equation expression formula is:

Predicted value is ${\mathrm{DT}}_2^{\left(0\right)}\left(21\right)=0.574173,$ ${\mathrm{DT}}_3^{\left(0\right)}\left(21\right)=0.374333,$ ${\mathrm{DT}}_4^{\left(0\right)}\left(21\right)=-0.473846,$ ${\mathrm{DT}}_5^{\left(0\right)}\left(21\right)=-0.432419,$ ${\mathrm{DT}}_6^{\left(0\right)}\left(21\right)=0.397095.$

The 6th step, can reach a conclusion in conjunction with Fig. 3 and predicted value: between single lattice at single lattice at transducer 2 places and transducer 3 places, rate of change between single lattice at single lattice at transducer 5 places and transducer 6 places is relatively more consistent, between single lattice at single lattice in transducer 4 places and transducer 5 places, rate of change between single lattice at single lattice at transducer 3 places and transducer 4 places is relatively more consistent, negative value represents that the temperature of single lattice at single lattice at transducer 5 places and transducer 4 places will be higher than the temperature of single lattice at single lattice at transducer 4 places and transducer 3 places, the state of cure (vulcanization) that pole plate between single lattice at single lattice at transducer 2 places and transducer 3 places also is described is greater than the state of cure (vulcanization) of pole plate between single lattice at single lattice at transducer 3 places and transducer 4 places, same reason, the state of cure (vulcanization) of pole plate is greater than the state of cure (vulcanization) of pole plate between single lattice at single lattice at transducer 4 places and transducer 5 places between single lattice at single lattice at transducer 3 places and transducer 4 places; Between single lattice at single lattice at transducer 1 place and transducer 2 places, rate of change is outstanding than single lattice at other transducer places, plate vulcanizing relatively serious (it is higher to vulcanize serious local temperature rise) is described between single lattice at single lattice at transducer 1 place and transducer 2 places, in general, the difference of single lattice rate of change at each transducer place does not reach the requirement of consistency set point 0.01, thereby repair process continues from the first step.

Whole repair process effect as shown in Figure 2, it is 48 hours that DK-2008 sets repair time, obviously shorten repair time as can be seen from Figure 2 of the present invention under the prerequisite that guarantees repairing effect, the extraordinary appearance that has suppressed to repair excessively phenomenon, repair process is to stop after 38 hours, and the DK-2008 repair process is repaired substantially when closing on 43 hours and is finished, but the capacity bust that occurs when continuing to repair, there was the reparation problem in expression.

The above only is preferred embodiment of the present invention, is not to be the restriction of the present invention being made other form, and any those skilled in the art may utilize the technology contents of above-mentioned announcement to be changed or be modified as the equivalent embodiment of equivalent variations.But every technical solution of the present invention content that do not break away to any simple modification, equivalent variations and remodeling that above embodiment did, still belongs to the protection range of technical solution of the present invention according to technical spirit of the present invention.

Claims (1)

1. based on the lead acid accumulator restorative procedure of time series tracking prediction, it is characterized in that: step is as follows

1.1 each single Ge WenduT of lead acid accumulator in the fixed cycle continuous sampling repair process _i ⁽⁰⁾(k);

Wherein: represent the single lattice of lead acid accumulator i, i ∈ [1,6], k are sampling instant;

Change DT 1.2 calculate the temperature rise of pole plate between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k), embodying formula is:

${\mathrm{DT}}_i^{\left(0\right)}\left(k\right)=T_i^{\left(0\right)}\left(k\right)-{T_{i-1}}^{\left(0\right)}\left(k\right),i\&GreaterEqual;2$

Change DT 1.3 set up the temperature rise of pole plate between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k) p rank autoregression model AR (p), the formula of embodying is:

ε wherein _kBe the white Gaussian noise sequence, p value maximum is got number of samples M one time, Be the model following coefficient;

1.4 the temperature rise of pole plate changes DT between the adjacent single lattice of lead acid accumulator _i ⁽⁰⁾(k) red pond criterion AIC is followed in the model order optimization of p rank autoregression model AR (p), optimization method, and the formula of embodying is:

AIC(p)＝ln(σ ²(p))+2p/M

Wherein, σ ²(p) be p rank model of fit residual error variances;

If there is p ₀∈ [1, p] makes $\mathrm{AIC}\left(p_0\right)=\underset{1\&le;p\&le;M}{\min }\mathrm{AIC}\left(p\right),$ P then ₀Be best autoregression model exponent number;

1.5 the temperature rise based on pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) p ₀Rank autoregression model AR (p ₀), the temperature rise of carrying out pole plate between the adjacent single lattice of lead acid accumulator changes DT _i ⁽⁰⁾(k) tracking prediction in advance, the formula of embodying is:

1.6 determine each single lattice pole plate temperature rise DT of lead acid accumulator _i ⁽⁰⁾Whether predicted value (k) is consistent, promptly

Whether equate,, otherwise get back to step 1.1 if equate then to stop repair process.

Images