WO2004070909A1

WO2004070909A1 - Pulse generation device for charging a valve-regulated lead-acid battery

Info

Publication number: WO2004070909A1
Application number: PCT/AU2004/000121
Authority: WO
Inventors: Lan Trieu Lam; Christopher Gerard Phyland; Nigel Peter Haigh; David Anthony Rand
Original assignee: Commonwealth Scientific And Industrial Research Organisation; International Lead Zinc Research Organization, Inc
Priority date: 2003-02-03
Filing date: 2004-02-03
Publication date: 2004-08-19

Abstract

There is disclosed a pulse generation device (1) for improving the life of a valve-regulated lead-acid battery (2), comprising a pulse generator capable of generating direct current charging pulses at a plurality of different frequencies corresponding to different ones of a plurality of potential sulfation patterns on the plates of lead-acid batteries, and configuration means (6) for configuring said pulse generator to generate pulses at a selected one of said plurality of frequencies, whereby said frequency can be selected on the basis of the expected sulfation pattern of the plates of the battery to which said pulse generation device is to be connected.

Description

PULSE GENERATION DEVICE FOR CHARGING A VALVE-REGULATED LEAD-ACID BATTERY

5 Related Art

The present application is based on_. and claims benefit of, US provisional application 60/444,559 filed 3 February 2003.

10 Field of the Invention

The present invention relates to a pulse generation device and method for improving the life of a lead-acid battery.

15

Background to the Invention

The proposed 42 -V powernet in automobiles requires batteries to provide a large number of shallow

20 discharge-charge cycles at a high rate. High-rate discharge is necessary for engine cranking, while high- rate charge is associated with regenerative braking. Batteries will therefore operate at these high rates in a partial state-of-charge condition — ^ΛHRPSoC duty' .

25

Under simulated HRPSoC duty, it has been found that the lead-acid batteries fail prematurely due to the progressive accumulation of "hard" lead sulfate mainly on the surf ces of the negative plates . The hard lead

30 sulfate is lead sulfate which is difficult to recharge because the lead sulfate layer cannot be converted efficiently back to sponge lead during charging either from the engine or from the regenerative braking. Eventually, this layer of lead sulfate develops to such

35 extent that the effective surface area of the plate is reduced markedly and the plate can no longer deliver the high-cranking current demanded by the automobile. This effect can impair battery performance and life, and can reduce charge-acceptance during regenerative braking. Moreover, the problem of over-sulfation is exacerbated during prolonged parking of the vehicle.

Accordingly, it would be desirable to improve the life of such batteries.

Summary of the Invention

The invention may be said to reside in a pulse generation device for improving the life of a lead-acid battery, comprising: a pulse generator capable of generating charging pulses at a plurality of different frequencies corresponding to different ones of a plurality of potential sulfation patterns on the plates of lead-acid batteries; and configuration means for configuring said pulse generator to generate pulses at a selected one of said plurality of frequencies, whereby said frequency can be selected on the basis of the expected sulfation pattern of the plates of the battery to which said pulse generation device is to be connected.

The invention may also be said to reside in a method for improving the life of a lead-acid battery, comprising: applying charging pulses to a lead-acid battery at a frequency corresponding to the expected sulfation pattern of the plates of said lead-acid battery.

The invention may also be said to reside in a method of charging a valve-regulated lead-acid battery operating under High-rate partial state of charge duty comprising applying direct current charging pulses having a frequency in the range of 50 kHz to 150 kHz to said battery.

The invention may also be said to reside in a method of charging a valve-regulated lead-acid battery operating under low-rate partial state of charge duty comprising applying direct current charging pulses having a frequency in the range of 10 to 500 Hz.

The invention may also be said to reside in a method of charging a valve-regulated lead-acid battery having a high carbon content and operating under high-rate partial state of charge duty comprising applying direct current charging pulses having a frequency in the range of 10 to 500 Hz.

The invention may also be said to reside in a pulse generation device for improving the life of a lead- acid battery, comprising: a pulse generator capable of generating direct current charging pulses at a plurality of different frequencies corresponding to different ones of a plurality of potential sulfation patterns on the plates of lead-acid batteries, said pulse generator being configured to generate pulses at at least two different frequencies corresponding to different patterns during at least an initial charging period.

The invention may also be said to reside in a method for improving the life of a valve-regulated lead- acid battery, comprising: applying charging pulses to a lead-acid battery at a at least two different frequencies corresponding to at least two possible sulfation patterns of the plates of said lead-acid battery during at least an initial charging period. Brief Description of the Drawings

A preferred embodiment of the invention will now be described in relation to the following drawings in which:

Figure la) is a schematic diagram of a possible pulse generation device arrangement;

Figure lb) is a schematic diagram of a pulse generation device;

Figure 2 is a solubility curve for the lead sulfate in sulfuric acid;

Figure 3 is a schematic representation of the distribution of lead sulfate in a negative plate subjected to low or high rate discharge;

Figure 4 is a schematic diagram representing the charging process of a negative plate after high rate discharge;

Figure 5 shows a 42 -V profile;

Figure 6 is a graph representative of cycle number VRLA batteries charged using a number of different schemes;

Figure 7 is a performance of VRLA battery VR10;

Figure 8 shows the performance of battery VR11;

Figure 9 shows the performance of battery VR12;

Figure 10 is a graph for a larger number of batteries under different pulse conditions; Figure 11 is an equivalent circuit for the plates of a lead acid battery;

Figure 12 is a schematic diagram showing electrical charges delivered under different pulse conditions;

Figure 13 is a graph of battery performance operated under different charging regimes;

Figure 14 shows a cycling regime;

Figure 15 shows the performance of three flooded batteries operating under the regime of Figure 14;

Figure 16 shows the performance of four flooded batteries; and

Figure 17 shows the performance of various high carbon content batteries .

Description of the Preferred Embodiment

The pulse generation technique of the preferred embodiment provides a pulse generation device for improving the life of a lead-acid battery. The inventors have determined that difference sulfation patters occur on their construction and the use to which they are put. Accordingly, the preferred embodiment provides a pulse generator 1 which is capable of generating charging pulses at a plurality of different frequencies which correspond to different potential sulfation patterns on the battery. That is, a single pulse generation device can be used with a number of different battery types and configured in accordance with the expected sulfation pattern. The pulse generation technique is particularly suited to valve- regulated lead-acid battery and the particular frequency and on times which have been developed for such batteries minimise the development of hard sulfate during both HRPSoC duty and stand conditions .

The technique involves the application of sets of charging pulses to the battery using a pulse generation device. A typical set-up is shown schematically in Figure 1. The pulse generator 1, delivers pulses of appropriate frequency and on-time to battery 2. The pulse generation device draws power from a power source 3. A programming signal 4, is used to configure the pulse generation device to produce a frequency corresponding to the expected sulfation pattern of battery 2. Depending on the configuration of the vehicle in which the technique is employed, the pulse generation device could be powered by an integrated starter and generator (ISG) , a small solar panel, or a small supercapacitor. In the last-mentioned option, the supercapacitor may be kept predominantly at a full SoC, either by the ISG or by regenerative braking during vehicle running, and is sized to provide continuous power for operation of the pulse-generation device even when the vehicle is parked for periods of several days.

In the first embodiment a pulse generator is provided by the circuit as shown in Figure 1 (b) . In this embodiment, a high voltage supply ac supply 5 is rectified and charges a large capacitor CI to produce a high-voltage dc supply of approximately 150V. This capacitor, in turn, charges the pulse capacitor C2 via the resistor RI. When the IGBT (Ql) is switched on by the microprocessor 6, the pulse capacitor C2 discharges via the current limit resistor R2 and provides a current to the battery with amplitude of:

r _χ V C2 ' BATTERY > pulse ~ no This assumes that the pulse on-time is relatively short so that the pulse capacitor (C2) does not significantly discharge. Resistor RI is used to limit the current drain on the rectifier and filter components during the pulse generation.

The microprocessor 6 is programmable to deliver the required pulse width and repetition rate of the pulses and thus provides configuration means for configuring the pulses generated by the generator to analog to digital converter 7. The microprocessor 6 can also monitor the battery voltage via the Analogue and use the rates of rise and fall of the battery voltage during charging and discharging to determine the appropriate pulsing parameters - e.g. frequency.

The microprocessor is configured to deliver low-frequency pulses in the frequency range of 10 Hz to 500 Hz and high-frequency pulses in the frequency range of

50 kHz to 150 kHz. Depending on frequency, the high-frequency pulses have an on-time of 1 to lOμs and an average current of 20 to 60 mA.

It will be apparent to persons skilled in the art that any number of different circuits may be employed. In one alternative embodiment, a boost converter has been successfully employed to deliver the current pulses. In this alternative embodiment, a buck converter is used to supply the energy to the boost converter. Both the buck converter and the boost converter are controlled by the microprocessor. This allows the microprocessor to control the pulse amplitude (via the buck circuit) and the pulse width (via the boost circuit) .

It will be apparent that while the apparatus of the present invention is highly desirable because it can be configured to a number of different batteries, the novel charging technique for VLRA batteries could be implemented in a dedicated pulse generation device which only applies charges at the appropriate frequency for such batteries.

The inventors have determined the following failure mechanism for lead-acid batteries. The discharge and charge processes at the negative plate can be expressed by reactions 1 to 4. During discharge, the conversion of sponge lead to lead sulfate proceeds via two steps. First, the sponge lead at the negative plate reacts with HSO₄- to form Pb²⁺, S0₄ ^2~ and H⁺, i.e., the so- called 'dissolution process¹ (reaction 1). Then, the Pb²⁺ combines with S0₄ ^2~ to form PbS0₄, i.e., the so-called

'deposition process' or 'precipitation process' (reaction 2) . The first step is an electrochemical reaction and thus involves electron transfer. Such transfer of electrons takes place only on the conductive sites, i.e., on fresh lead. The rate of the electrochemical reaction is therefore dependent not only on the diffusion of HS0₄- species, but also on the effective surface-area of the sponge lead. On the other hand, the second step is a chemical reaction and proceeds with a rate which is acid dependent. The solubility of lead sulfate does not increase with increase in sulfuric acid concentration. Rather, it reaches a maximum value at 10 wt.% sulfuric acid (1.07 rel.dens.), and then decreases rapidly with further increase in concentration (Fig. 2) . Thus, the Pb²⁺ will precipitate as lead sulfate at concentrations above the solubility curve. Clearly, for a given concentration of Pb²⁺ above ~1 mg l^"1, the deposition (or precipitation) of Pb²⁺ to lead sulfate will be faster at plate locations which experience high concentrations of acid. During the initial stages of the discharge of a fully-charged negative plate, electron transfer can take place at any location because the entire plate is conductive. Accordingly, the discharge process (both dissolution and deposition steps) occurs — both on the surfaces and in the interior of the plate. Nevertheless, the reaction in the interior of the plate will soon slow down and/or stop, while that at the surfaces of the plate will continue to proceed. This is because less acid is available in the interior.

Discharge process

Dissolution Pb * HSO4- - 2e ^■*^* Pb²⁺ + SO₄ ^2" + H* (1)

1 Deposition

(2)

Charge process

Dissolution PbSO₄ ~_\ »~ Pb²⁺ + SO₄ ²- + 2e~ + H* (³)

' ^► Pb + HSO4- (4)

Deposition

The depth to which lead sulfate penetrates is dependent on the rate of discharge, as well as on the density and surface area of the plate. Paste density is the key factor in providing the macropores (or 'avenues') which are necessary for the transport of solution and ionicspecies to and from the reaction sites within the interior of the plate, while surface area provides sites for the current-generating electrochemical reaction. For the same paste density and surface area, the extent to which lead sulfate can penetrate is determined by the discharge rate. Under low-rate discharge (i.e., < 0.4C), the dissolution rate of Pb²⁺ from each lead crystal is slow and, therefore, the accompanying consumption of HS0₄ ^_ in the interior of the plate is likely to be counterbalanced by the diffusion of HS0₄ ^" from the bulk of the electrolyte. Furthermore, the subsequent deposition of Pb²⁺ to PbS0₄ (reaction 2) also occurs slowly due to the low supersaturation of Pb²⁺ in the vicinity of each mother lead crystal. (Note that, the deposition rate of PbS0₄ is proportional to the degree of supersaturation of Pb²⁺ in the sulfuric acid solution, i.e., the higher the supersaturation, the faster is the deposition rate.) Since the rate of deposition (reaction 2) is slow, newly formed PbS0₄ tends to precipitate preferentially on the already-deposited PbS0₄ crystals, i.e., growth rate > nucleation rate. Consequently where there is a low-rate of discharge, the deposited lead sulfate will continue to grow to various sizes of discontinuous crystals 10, both on the surface 10a and in the interior 10b of the negative plate. This form of lead sulfate is particularly desirable on the surface of the plate, as it provides an open structure that facilitates the ingress of HS0₄ ^~ ions. Therefore, the discharge process can proceed deep into the interior of the plate. Accordingly, the lead sulfate develops in an even sulfation pattern throughout the cross-section of the negative plate (Fig. 3(a)).

The formation of lead sulfate is quite different under high-rate discharge, e.g., under cranking-current (~18C) conditions. The electrochemical reaction (i.e., reaction 2) now proceeds so rapidly that the diffusion rate of HSO₄ ^"" cannot catch up with the consumption rate. Consequently, lead sulfate forms mainly on the surface (10c, lOd) of the plate as shown in Figure 3b. Moreover high-rate discharge generates a very high supersaturation of Pb⁺ in the vicinity of each mother lead crystal. The lead sulfate will therefore precipitate rapidly on any available surface, irrespective of whether this be sponge lead 11 or already-deposited lead sulfate, i.e., nucleation rate > growth rate. Thus, a compact layer of tiny lead sulfate crystals will develop on the surface of - li ¬

the plate. This will reduce the effective surface area for electron transfer and will also hinder the diffusion of HS0₄ ^~ into the interior of the plate (Fig. 3(b)). Under such conditions, the discharge reaction cannot proceed into the interior, but stops at the surface of the plate and at the walls of the pores.

During charging, the conversion of lead sulfate to sponge lead also proceeds via two reactions, namely, dissolution and deposition. Nevertheless, the nature of each of these reactions differs from that of the corresponding discharge reactions. Dissolution is now the chemical reaction, while the subsequent deposition is the electrochemical reaction. The lead sulfate first dissociates to Pb²⁺ and S0₄ ^2~ ions. The Pb²⁺ then receives to two electrons and reduces to lead. Simultaneously, S0₄ ^2~ combines with H⁺ to form HS0₄ ^~. The electrons flow to the active sites in the negative-plate material via the grid members because the electrical resistance of the grid metal is much smaller than that of the discharged material. In addition to the reduction of Pb²⁺ to lead, there is the competing reaction of hydrogen evolution. In general, hydrogen evolution only takes place near the end of charging due to the fact that: (i) most of the PbS0₄ has been converted to lead and, correspondingly, the sulfuric acid concentration, i.e., H⁺ concentration, has increased; (ii) further dissolution of PbSCa to Pb²⁺ and Sθ²⁺ is slow. However, hydrogen can also be involved during the e&srly stages of the charging process, if the dissolution of PbS0₄ is hindered.

The recharge mechanisms are as follows. Firstly, recharge of the negative plate after it has been deeply discharged at a low rate occurs after, as mentioned above, lead sulfate is formed throughout the entire cross-section of the plate and the relative density of the acid after discharge is low because of the high utilization of the active material. The dissolution of PbS0₄ to form Pb²⁺ and S0₄ ^2_ increases at the low concentrations (see, Fig. 2) . Thus, the subsequent reduction of Pb²⁺ to sponge lead can take place smoothly before the evolution of hydrogen. With an overcharge factor of ~10%, the plate can be brought to a fully-charged state without any difficulty. This is also true when the plate is subjected to low-rate PSoC cycling with equal amounts of charge input and charge output. In such duty, the SoC of the negative plate decreases with cycling, but can be brought to 100% after the application of an equalization charge.

By contrast, the recharge of the negative plate after deep discharge at a high rate is difficult. Since high-rate discharge cannot proceed into the interior of the plate, but stops at the surface, the utilization of the active material is low. Consequently, the relative density of the acid after discharge is still at a high level and this decreases the dissolution of PbS0₄ (see, Fig. 3) . The lower concentration of Pb²⁺ will then impede the subsequent electrochemical reaction and, during the early stages of charging, will cause the negative-plate potential to become more negative to such extent that hydrogen can start to evolve. Furthermore, as mentioned above, the electrons flow from the grid members toward the surfaces of the plate. These electrons will reduce some hydrogen ions to hydrogen gas before reaching the lead sulfate layer (Fig. 4) . Thus, complete conversion of lead sulfate at the plate surface cannot be achieved, even with an overcharge of 10%, because of the combined effects of the early evolution of hydrogen and the oxygen- recombination reaction. Furthermore, the overcharge factor will increase with cycling because progressive water loss will dry-out the separator, increase the amount of oxygen reaching the plate, and hence will enhance the level of oxygen recombination. Thus, lead sulfate will accumulate on the surface of the negative plate and, eventually, the battery will be unable to provide sufficient power for engine cranking.

From the above, the inventors have determined that the cycleability of batteries can be increased if current can be directed to the battery in a manner consistent with the sulfation pattern and in particular, the cycleability of VRLA batteries under HRPSoC duty can be enhanced if, during charging, current can be concentrated on the surfaces of the negative plates.

When a direct current is passed through a conductive wire, the current will be distributed evenly throughout the entire cross-section of the wire. When, however, an alternating current (a.c.) or a direct current (d.c.) in pulsed form is passed through the same wire, the current will be localized on the perimeter of the conductor. This is termed the skin effect' as only the outer ^λ skin' of the cross-section is effectively carrying the current. The penetration of the current is called the ^λskin depth' and can be calculated from the following equation:

Skin depth = (p/π/μ)^1/2 (5)

Where: p = bulk resistivity of the current carrier (e.g., 2.053 x 10^"7 Ω-m for lead); μ = magnetic permeability of free space (1.257 x 10^"6 Wb per A-m) ; / = frequency of pulsed current (Hs) . In the case of a battery, the skin depth is degree to which the current penetrates into the interior of the plate.

Equation (5) shows clearly that the charging current will be concentrated more on the surfaces of the plate, i.e., on the lead sulfate layer, when high- frequency a.c. and/or d.c. pulses is/are used. Devices that provide direct-current pulses of low, medium or high-frequency were designed, constructed and used in tests on commercial VRLA batteries (12 V, C₂o or 20-h capacity = 33 Ah) . Initially, five batteries were prepared and subjected to repetitive sets of a 42 -V profile at 40°C (Fig. 5) . This profile is an accepted regime for evaluating the durability of VRLA batteries under HRPSoC duty during both charge 20 and discharge 21. The profile has a short duration (2.35 min) and is composed of several current steps that simulate the power requirements of the battery during vehicle operation, i.e., idle—stop 22, cranking 23, power assist 24, engine charging 25, and regenerative charging 26. The critical step is the cranking period over which the battery must deliver a current of 300 A for 0.5 s, i.e., a current equal to ~18C. Each application was considered to be 'one cycle' and a maximum of 1200 cycles was applied. The test was terminated when the batteries could not sustain at least 960 cycles (i.e., 0.8 x 1200 cycles) due to decrease in the end-of-discharge voltage to the cut-off value of 9.6 V during cycling. Otherwise, the batteries were charged fully and then subjected to a further set of 1200 cycles.

As shown in Figure 6, batteries without pulses

(VR1, VR2) or with (VR3 to VR5) low-frequency pulses (0.21 kHz) failed prematurely at between 10 000 and 10 650 cycles. Further analysis showed that failure was due to the progressive accumulation of lead sulfate mainly on the surface of the negative plates.

The latter pulses could also foe combined with low- frequency (0.21 kHz) or high-frequency d.c. pulses and operated only during the on-time of the d.c. pulses. Accordingly, further four VRLA batteries (i.e., VR6 to VR9) were prepared and subjected to the same test conditions as for batteries VRl to VR5 using pulse generators designed to provide d.c. pulses of higher frequency (i.e., 8.75 kHz) and a.c. ^λring' pulses of 1.05 MHz.. These batteries were superimposed with combined a.c. ^Λring' and low-/high-frequency pulses (VR6 - VR8) or only high-frequency pulses (VR9) .

As shown in Figure 6, the performance of VRLA batteries with either combined a.c. 'ring' and d.c. pulses or only high-frequency pulses is superior (15 000 to 18 800 cycles) to that of batteries without/with low- frequency pulses (10 000 to 10 650 cycles) . Although these batteries still failed because of negative-plate sulfation, build-up of lead sulfate occurred throughout the entire cross- ection of the plate rather than predominantly on the surface as found in batteries without/with low- frequency pulses. This indicates that the use of either combined a.c. 'ring' and d.c. pulses or only high-frequency d.c. pulses has assisted the charging of lead sulfate on the plate surface, and has allowed the discharge process to penetrate deep into the interior of the plate. Thus, the performance of the batteries is enhanced.

Table 1. Pulse conditions for further evaluation of battery performance.

A further series of tests were conducted to optimize only the d.c. pulse technique because this technique is easier to control than the a.c. 'ring' alternative. Accordingly, batteries (VR10 to VR12) were prepared and subjected to the 42 -V profile, but with the new pulsed conditions listed in Table 1. These pulses have a frequency which is higher than the previously examined d.c. pulses.

The performance of battery VR10 with d.c. pulses of on-time 0.6 μs, frequency 87.5 kHz and average pulsed current 20 mA is shown in Fig. 7. The battery completed 8 sets of the profile but on the ninth set, the battery voltage reached the cut-off value of 9.6 V at the 900th cycle. Thus, the total service given by this battery is 10 500 cycles, which is similar to that of batteries without pulses (see Fig. 6) . The superimposition of pulses with a high average pulsed current (60 mA) gave an improved performance, viz., 14 000 cycles for battery

VRll, Fig. 8. The increased average current was achieved by raising the amplitude of the pulse current from 400 to 1200 mA, but keeping the on-time and frequency unchanged. A further, and remarkable, increase in cycle-life — 32 000 cycles — was obtained from battery VR12 with pulses of the same average current as that for battery VRll (Fig. 9) , except that this current was achieved by increasing the on-time from 0.6 to 1.8 μs while keeping the pulsed current and frequency as same as that of battery VR10 (i.e., 400 mA and 67.5 kHz). The life performance of all VRLA batteries examined to date is summarized in Fig. 10.

After reaching the cut-off voltage of 9.6 V, battery VR12 was subjected to further sets of cycling, but with a new cut-off voltage of 7.2 V. This further cycling was conducted in order to determine the practical life of the battery as the cut-off voltage used to evaluate batteries under cranking-rate discharge is generally set at 7.2 V, rather than 9.6 V. It can be seen that the battery still performed well and provided another 11 sets of cycling (i.e., total cycle number = 45 200 cycles), but unfortunately failed rapidly afterwards . Examination of the change in voltage, charge-to- discharge ratio and internal resistance of the above three batteries during cycling revealed that there is little differences in the top-of-charge voltages (ToCVs) and charge-to-discharge ratios. On the other hand, the degree of increase in the internal resistance with the application of a cycling set is smaller in batteries VRll and VR12 with high average pulsed current (60 mA) than in battery VR10 with low average pulsed current (20 mA) . Battery VR12 shows the slowest increase in internal resistance and, therefore, also displays the slowest decrease in the end-of-discharge voltage (EoDV) with cycling.

We propose that the lead sulfate layer can be 'charged' by localizing the current on the surface of the plate via the use of d.c. pulses of high-frequency — the higher the frequency, the greater is the concentration of current on the plate surfaces. This is known as the skin effect, and is applicable to both conductive wires and battery plates. A pure resistance can be considered for a conductive wire, but not for a battery plate. A simple equivalent circuit for a positive or a negative plate in a battery is presented in Fig. 11. Apart from different values for the parameters, the basic model for both electrodes is considered to be the same. The equivalent circuit is composed of an inductance L, a contact resistance R_c, a capacitance C__ι , a Faradaic resistance R_f, and a solution resistance R_so . The inductance L is simply- caused foy the metallic connection either between the cable and the battery or between the terminals, bus-bar and plate lugs. The contact resistance R_c arises from this connection as well as the conductivity of the plate. Thus, the summation of R_c and R_soι is the internal resistance of the battery. The capacitance C_u is developed by the double layer at the interface between the plate and the electrolyte solution. Finally, R_f is a nonlinear resistance which represents the electrochemical reaction, i.e., conversion of lead sulfate to lead. Clearly, the charging efficiency will be increased if more charge is delivered to the R_f component.

The above model can be used to explain the differences in cycle-life performance of batteries VR10, VRll and VR12, even though these batteries were cycled under the superimposition of pulses with the same frequency, but of different pulsed current, on-time and off-time (see Table 1) . As the pulse frequency increases, the inductance component will produce a greater back electromotive force (back 'e f') to the pulses. Consequently, there will be a greater energy loss when d.c. pulses of high-frequency pass through the inductance. (Note, it is understood that the pulsed current will become distorted in profile when passed through the inductance, but for simplicity, it is shown as a square profile in Fig. 11) .

Since the three batteries are subjected to pulses of the same frequency, the respective energy losses will be similar. After passing through the contact resistance and entering the parallel circuit, the bulk of the pulsed current, I_plr first goes to charge the capacitor. When the capacitor C__\ι is charged and the potential of the plate increases, the charging current to the capacitor will decrease to a very low value. The current, which is driven hy the increase in plate potential, can now pass through the resistance R_f and will progressively increase. (It should be noted that in order to pass the current through this resistance, a certain driving force, i.e., overpotential, is required.) This indicates that the final part, not the early part, of the current is beneficial for the conversion of lead sulfate. Although the discharge from the double-layer capacitor to the Faradaic resistor occurs during the initial off-time of the pulse, it's effect on the deposition reaction (e.g., conversion of lead sulfate to sponge lead) has been shown to be negligible.

Figure 12 shows I_p fpr VR10-VR12 (Figures 12(a) to 12(c) respectively). In each case, I_p 30 is split into I_pi 301 and I_p2 302. For battery VR10, although more current may be concentrated on the surface of the negative plate through the use of high-frequency pulses (see Table 1) , the energy loss in the inductance and capacitor does not allow sufficient charge to be passed through the Rf component to break down all of the lead sulfate crystals. Thus, the performance of this battery does not improve with pulsing. For batteries VRll and VR12, substantial charge is passed through the Rf component because of the use of a high average pulsed-current and, therefore, cycle-life is improved, particularly in the case of battery VR12 which sustained 32 000 cycles. For these two batteries, the average pulsed-current 30 is raised from 20 to 60 mA by a three- fold increase of either the pulsed current 30b (battery VRll) or the on-time 30c (battery VR12) . Thus, the electrical quantity used to charge the C_dl component in each pulse is similar to that provided by the pulses superimposed on the battery VR10 but the electrical charge passed through the R is three times greater (Fig. 12) , and, accordingly, improved cycle-life is obtained. The significant difference in cycling improvement between batteries VRll and VR12 (14 000 vs. 32 000 cycles) is, hot^ever, surprising. It has been observed that the current transients of the pulses produce an initial decaying oscillation. This oscillation lasts for about 500 ns. The oscillation proportion of the current pulse is probably ineffective for the conversion of lead sulfate, accordingly a longer on time or a cleaner square pulse with reduced oscillation improves changing performance . In order to confirm the performance of the high- frequency pulse (i.e., f = 87.5 kHz, t_on = 1.8 μS) , two commercial batteries were prepared and subjected to the same test as shown in Fig. 5. As the capacity of these batteries was smaller than that of the previous batteries, the amplitude of each current step was scaled down by a factor of 4.3. As shown in Figure 13, battery VR 13 which was cycled without pulse charging failed at 21 000 cycles, while battery VR 14 which was charged with high-frequency pulses enjoyed a much longer life of 42 000 cycles.

A further series of tests were conducted to examine the effects of pulse frequency when the batteries were cycled under partial state-of-charge, but at low rate. Accordingly, several commercial, flooded, 12 -V batteries (C₂o = 35 Ah, reserve capacity = 40 min) were used in the screening test which was conducted at room temperature. Prior to the test, the batteries were brought to full charge by applying a constant-current (2.5 A max. ) -constant-voltage (15.5 V) procedure for a total of 24 h. After charging, the batteries were operated successively through a set of four different PSoC windows, namely, 90-60% 30, 70-40% 31, 80-40% 32 and 90-40% 33 as shown in Figure 14. During each PSoC window, the batteries were discharged at 25 A (i.e., 1.36C, C = 1-h capacity) and recharged at 10 A (i.e., 0.54C) with equal amounts of charge input and charge output for a maximum of 30 cycles. No equalisation charge was applied between each PSoC window. When the voltage of each battery reached the cut-off voltage of 7.2 V during cycling, the battery was fully charged and subjected to a further three sets of PSoC windows. During cycling, the end-of- discharge and end-of-charge voltages of each battery were recorded.

Three flooded, 12-V batteries were cycled to the cut-off voltage of 7.2 V. Battery FE1 was cycled without pulses, while batteries FE2 41 and FE3 42 were cycled with pulses of different frequency but of similar average pulsed current 20 mA. The pulse conditions for these latter two batteries are given in Table 1.

Table 2. Pulse conditions.

The performance of batteries FE1 40 to FE3 42 is presented in Fig. 15. During each set of PSoC windows, battery FE2 41, which was fitted with pulse device of 217 Hz, sustains the most cumulative cycles after four sets of PSoC windows. Battery FE2 42 with 9.09 kHz device shows only a slight improvement in cycle performance over battery FE1 40 without pulses. In order to confirm the consistency of these results, the same test was repeated with a further two batteries: FE4 43 without pulses and FE5 44 with pulses of low frequency. The performance of the batteries is shown in Fig. 16. Clearly, there is little difference in the performance of each battery within a corresponding pair, namely, FE4 43, FE1 40 (without pulses) and FE5 44, FE2 41 (with pulses) . This indicates that the results are reproducible. As mentioned previously, the lead sulfate develops evenly across the cross-section of the plates, when the batteries were cycled under partial state-of-charge and at low rate (see Fig. 3(a)) . Since the use of low-frequency pulses can distribute the pulse current evenly across the cross section of the plate, the batteries using low-frequency pulses give better service life. For further confirmation of the effects of low- frequency pulses, a set of small VRLA batteries were prepared and subjected to the 42 -V profile as shown in Fig. 5, but with reduced amplitude of each current step by a factor of 4.3. These batteries were specially designed and built for hybrid-electric vehicle application. The main difference to conventional batteries is that the negative plates of these batteries have a much higher carbon content than conventional batteries (i.e., 1.0 wt.% vs. 0.2 wt.% in conventional designs) . The high carbon content leads to even distribution of the development of lead sulfate during high rate discharge evenly across the cross-section of the negative plate (rather than concentrate on surf ce) .

The performance of these HEV batteries is shown in Fig. 17. Batteries VR15 to VR18 were cycled without pulses, while batteries VR19 to VR21 were with pulses of different frequencies. There is significant variation in cycle-lives of batteries without pulses being in the range between 27 000 cycles and 65 000 cycles. The cycle-lives of Battery VR16 and 17 are very close being about 27 000 cycles, while that of batteries VR15 and VR18 are 65 000 and 56 000 cycles, respectively. Batteries VR19 and VR20 were cycled with pulses. Battery VR19 with high-frequency pulse of 87.5 kHz failed at 26 000 cycles. The cycle-life of battery VR20 was further decreased to 19 000 cycles when the battery was fitted with pulses of higher frequency (i.e., 148 kHs) . Battery VR21 was cycled with low-frequency pulses (i.e., 217 Hz). As expected, the battery VR21 gave 98 000 cycles and this cycling performance is superior than those of batteries without pulses. These results show clearly that the beneficial effect of pulse frequency on the cycle-life performance of batteries is dependent upon the distribution of lead sulfate formed during discharge. The high-frequency pulse is effective when the lead sulfate is formed on the surfaces of the negative plates (e.g., under high-rate discharge and batteries with low concentration of carbon) . On the other hand, the low-frequency pulses are effective when lead-sulfate is formed evenly across the cross- section of negative plates e.g., under low-rate discharge or high-rate discharge, but with batteries of high carbon content .

It will be apparent from the foregoing description that it is advantageous to provide a device which can be configured to the expected sulfation pattern

- i.e. one which can be pre-configured or set. However, it is also possible for a device which is self-configuring

- i.e. which determines from the batteries changing behaviour what type of pulses should be applied.

During an initial charging period, the pulse generator is configured to generating pulses consisting of a combination of high-frequency and low-frequency pulses. The pulse generator is configured to vary the ratio of high-frequency pulses x to low-frequency pulses y (e.g., r = x:y) , based upon the rates of rise and fall of battery voltage during charging and discharging.

For example, if the rates of rise and fall of the battery voltage during charging and discharging are high, y is set as zero after the initial charging period (i.e. only high-f equency pulses) or x is set to be much greater than y (i.e. the number of high-frequency pulses are more than that of low-frequency pulses.

Similarly, if the rates of rise and fall of the battery voltage during charging and discharging are low, x is set as zero (only low- frequency pulses) or y is set greater than x (the number of low-frequency pulses are more than that of high-frequency pulses) . Similarly, if the rates of rise and fall of the battery voltage during charging and discharging are stable, x and y are set to be the same (i.e. the number of high-frequency pulses are equal to that of low- frequency pulses.

It will be apparent to persons skilled in the art that various modifications may be made without departing from the scope of the invention described herein.

Claims

CLAIMS :

1. A pulse generation device for improving the life of a valve-regulated lead-acid battery, comprising: a pulse generator capable of generating direct current charging pulses at a plurality of different frequencies corresponding to different ones of a plurality of potential sulfation patterns on the plates of lead-acid batteries; and configuration means for configuring said pulse generator to generate pulses at a selected one of said plurality of frequencies, whereby said frequency can be selected on the basis of the expected sulfation pattern of the plates of the battery to which said pulse generation device is to be connected.

2. A pulse generation device as claimed in claim 1, wherein said pulse generator is capable of generating pulses at a first frequency corresponding to a sulfation pattern where lead sulfate is concentrated at the surfaces of the plate and at a second frequency corresponding to a sulfation pattern where lead sulfate is more evenly distributed.

3. A pulse generation device as claimed in claim 2, wherein said first frequency is higher than said second f equency.

4. A pulse generation device as claimed in claim 2, wherein said first frequency is in the range of 50 kHa to

150 kHz and said second frequency is in the range of 10 Hz to 500 Hz.

5. A pulse generation device as claimed in claim 1, wherein said pulses are square wave pulses.

6. A pulse generation device as claimed in claim 2, wherein the first frequency pulses have an on time in the range of 1 to 10 μs and an average current in the range of 20mA to 60 mA.

7. A pulse generation device as claimed in claim 1, wherein said pulse generation device is configured on the basis of the expected rate of charge/discharge to thereby determine the expected sulfation pattern.

8. A pulse generation device as claimed in claim 1, wherein said pulse generation device is configured on the basis of the charging behaviour of the battery.

9. A method for improving the lif of a valve- regulated lead-acid battery, comprising: applying charging pulses to a lead-acid battery at a frequency corresponding to the expected sulfation pattern of the plates of said lead-acid battery.

10. A method as claimed in claim 9, comprising: providing a pulse generator capable of generating direct current charging pulses at a plurality of different frequencies corresponding to different ones of a plurality of potential sulfation patterns on the plates of lead-acid batteries; and configuring said pulse generator to apply charging pulses to said battery on the basis of the expected sulfation pattern of the plates of the battery.

11. A method as claimed in claim 9, further comprising determining the sulfation pattern from the expected rate of charge/discharge of the battery.

12. A method of charging a valve-regulated lead-acid battery operating under High-rate partial state of charge duty comprising applying direct current charging pulses having a frequency in the range of 50 kHz to 150 kHz to said battery.

13. A method as claimed in claim 12, wherein said charging pulses have an on time of lμs to lOμs and an average current in the range of 20mA to 60mA.

14. A method of charging a valve-regulated lead-acid battery operating under low-rate partial state of charge duty comprising applying direct current charging pulses having a frequency in the range of 10 to 500 Hz.

15. A method of charging a valve-regulated lead-acid battery having a high carbon content and operating under high-rate partial state of charge duty comprising applying direct current charging pulses having a frequency in the range of 10 to 500 Hz.

16. A pulse generation device for improving the life of a lead-acid battery, comprising: a pulse generator capable of generating direct current charging pulses at a plurality of different frequencies corresponding to different ones of a plurality of potential sulfation patterns on the plates of lead-acid batteries, said pulse generator being configured to generate pulses at at least two different frequencies corresponding to different patterns during at least an initial charging period.

17. A pulse generation device as claimed in claim 16, wherein said pulse generator is capable of generating pulses at a first frequency corresponding to a sulfation pattern where lead sulfate is concentrated at the edges of the plate and at a second frequency corresponding to a sulfation pattern where lead sulfate is more evenly distributed.

18. A pulse generation device as claimed in claim 17, wherein said first frequency is higher than said second frequency.

19. A pulse generation device as claimed in claim 17, wherein said first frequency is in the range of 50kHz to

150kHz and said second frequency is in the range of 10Hz to 500Hz.

20. A pulse generation device as claimed in claim 16, wherein said pulses are square wave pulses.

21. A pulse generation device as claimed in claim 17, wherein the first frequency pulses have an on time in the range of greater Iμs to lOμs and an average current in the range of 20mA to 60mA.

22. A pulse generation device as claimed in claim 16, wherein said device has means to determine the charging behaviour of said battery during said initial period and configuration means to configure the battery to generate pulses corresponding to the charging behaviour of the battery after the initial period.

23. A pulse generation device as claimed in claim 16, wherein said configuration means is configured to alter the ratio of high-frequency and low frequency pulses on the basis of the rate of rise and fall of the battery voltage during charging.

24. A method for improving the life of a valve-regulated lead-acid battery, comprising: applying charging pulses to a lead-acid battery at a at least two different frequencies corresponding to at least two possible sulfation patterns of the plates of said lead-acid battery during at least an initial charging period.

25. A method as claimed in claim 24, comprising: providing a pulse generator capable of generating charging pulses at a plurality of different frequencies corresponding to different ones of a plurality of potential sulfation patterns on the plates of lead-acid batteries; and configuring said pulse generator to apply charging pulses to said battery at at least two different frequencies corresponding to different sulfation patterns during at least an initial charging period.

26. A method as claimed in claim 24, further comprising adjusting the pulse applied to the battery on the basis of the charging behaviour of the battery.