Parallel DC power sources with different characteristics

[0002] 1. Technical Field

[0003] The present invention relates to a method and structure for electric power management that economically satisfies peak power demand and is able to immediately satisfy a sudden substantial increase in power demand.

[0004] 2. Related Art

[0005] A power management system having an energy source may be used to supply electrical power to a facility such as a residence, a factory, a shopping center, etc. Power demand by such a facility may vary greatly with time. For example, the power demand may be very high for short periods of time, and relatively low for long periods of time. The peak power load can be several times an average power load. The energy source must be capable of satisfying the peak power demand. The fabrication costs and maintenance costs of the power management system generally increase as the system's rated peak power increases.

[0006] Additionally, some energy sources are unable to immediately satisfy a sudden substantial increase in power demand. For example, a fuel cell requires hydrogen gas for generating power. Thus a sudden increase in power demand requires a corresponding increase in available hydrogen gas. Since it is often not possible to generate hydrogen gas sufficiently rapidly to keep up with the sudden substantial increase in demand, the fuel cell voltages may fall below acceptable limits and consequently suffer damage.

[0007] There is a need for an economical method and structure for satisfying peak power demand, and also for accommodating a sudden substantial increase in power demand.

[0008] The present invention provides a power management system for supplying electric power to a load, comprising:

[0009] a first electrical circuit element that includes a first direct current (DC) energy source in series with a first rectifier that prevents current backflow into the first energy source; and

[0010] a second electrical circuit element that includes a second DC energy source in series with a second rectifier that prevents current backflow into the second energy source, wherein the second electrical circuit element is in parallel with the first electrical circuit element, and wherein an open circuit voltage V₁₀of the first energy source exceeds an open circuit voltage V₂₀of the second energy source.

[0011] The present invention provides a method for supplying electric power to a load, comprising:

[0012] forming a first electrical circuit element that includes a first direct current (DC) energy source in series with a first rectifier that prevents current backflow into the first energy source; and

[0013] forming a second electrical circuit element that includes a second DC energy source in series with a second rectifier that prevents current backflow into the second energy source, wherein the second electrical circuit element is in parallel with the first electrical circuit element, and wherein an open circuit voltage V₁₀of the first energy source exceeds an open circuit voltage V₂₀of the second energy source.

[0014] The present invention provides an economical method and structure for satisfying peak power demand, and also for accommodating a sudden substantial increase in power demand.

[0015] FIG. 1 depicts a schematic diagram of a power management system for supplying electric power to a load, comprising a first circuit element having a first direct current energy source and a first diode in series, in parallel with a second circuit element having a second energy source and a second diode in series, and in parallel with the load, in accordance with embodiments of the present invention.

[0016] FIG. 2 depicts FIG. 1 with the second energy source being a battery, and after a battery charger has been coupled to the battery.

[0017] FIG. 3 depicts FIG. 2 after the second diode has been replaced by a silicon controlled rectifier.

[0018] FIG. 4 depicts FIG. 3 after a DC to DC converter has been added in series with the battery.

[0019] FIG. 5 depicts FIG. 4 after addition of a signal processor for transmitting a signal to the DC to DC converter to dynamically control an output current of the first energy source.

[0020] FIG. 6 depicts a steady-state voltage versus current characteristic curve of the first energy source of FIG. 1 after subtracting a voltage drop across the first diode.

[0021] FIG. 7 depicts a steady-state voltage versus current characteristic curve of the second energy source of FIG. 1 after subtracting a voltage drop across the second diode.

[0022] FIG. 8 depicts a steady-state load voltage drop versus load current characteristic curve for the circuit of FIG. 1, based on the characteristic curves of FIGS. 6 and 7.

[0023] FIG. 9 depicts the steady-state load voltage drop versus load current characteristic curve of FIG. 8 with an addition of a transient operational path following an increase in load power.

[0024] FIG. 1 illustrates a schematic diagram of a power management system 5 for supplying electric power to a load 30, comprising a first circuit element 7 in parallel with a second circuit element 8, and the load 30 in parallel with each of the first circuit element 7 and the second circuit element 8.

[0025] The first circuit element 7 includes a first energy source 10 in series with a diode 32. The first energy source 10 is a first direct current (DC) energy source. The first energy source 10 may comprise a conventional energy source such as, inter alia, fossil fuel, natural gas, hydroelectric power, etc. Alternatively, the first energy source 10 may comprise an alternative energy source such as, inter alia, fuel cells, solar cells, wind power, and biomass power. If alternating current (AC) power is generated as a precursor to the first energy source 10, then the AC power is rectified to produce the first energy source 10 as DC power. The first circuit element 7 includes a circuit path starting at a node 60, then to the first energy source 10, and then to a node 61, as shown. An anode of the first diode 32 is conductively coupled to a positive terminal 11 of the first energy source 10. Alternatively, a cathode of the first diode 32 could have been conductively coupled to a negative terminal 12 of the first energy source 10. The positioning and orienting of the first diode 32 serves to prevent current backflow into the first energy source 10; i.e., the first diode 32 prevents electrical current from flowing into the first energy source 10 at the positive terminal 11, since such backflow of current into the first energy source 10 could damage the first energy source 10. Note that the first diode 32 could be replaced by any rectifying electronic device (i.e., any rectifier such as a silicon controlled rectifier) that prevents said backflow of current into the first energy source 10.

[0026] The second circuit element 8 includes a second energy source 14 in series with a second diode 33. It should be noted that the second energy source 14 of FIG. 1 is replaced by a battery 20 in FIGS. 2-6, discussed infra. The second energy source 14 is a second direct current (DC) energy source. The second energy source 14 may comprise any energy source capable of responding instantaneously to a sudden demand in power, such as, inter alia, a battery or a series of batteries. Any type of battery such as, inter alia, a lead-acid battery, may be used. If the first energy source 10 comprises fuel cells, an operating voltage range may include, inter alia, about 40-90 volts for the first energy source 10, and about 48-54 volts the second energy source 14. The second circuit element 8 includes a circuit path starting at the node 60, then to the second energy source 14, and then to the node 61, as shown. An anode of the second diode 33 is conductively coupled to a positive terminal 15 of the second energy source 14. Alternatively, a cathode of the second diode 33 could have been conductively coupled to a negative terminal 16 of the second energy source 14. The positioning and orienting of the second diode 33 serves to prevent current backflow into the second energy source 14; i.e., the second diode 33 prevents electrical current from flowing into the second energy source 14 at the positive terminal 15. A dangerously high backflow of current into the second energy source 14 could damage the second energy source 14. Such dangerously high backflow of current into the second energy source 14 could occur if a voltage output of the first energy source 10, after being decremented by a forward-biased voltage drop across the first diode 32, is higher than a maximum voltage across the second energy source 14. Note that the second diode 33 could be replaced by any rectifying electronic device (i.e., any rectifier such as a silicon controlled rectifier) that prevents said backflow of current into the second energy source 14.

[0027] The load 30 could be, inter alia, a DC load, or an inverter that converts DC to AC.

[0028] The following notation is used for describing how current is supplied to the load 30

[0029] V₁: denotes a voltage at the positive terminal 11 of the first energy source 10.

[0030] I₁: denotes a current at the positive terminal 11 of the first energy source 10.

[0031] P₁: denotes a power output from the first energy source 10.

[0032] V₁₀: denotes an open circuit voltage of V₁(i.e., V₁when I₁=0).

[0033] R₁: denotes an equivalent internal resistance of the first energy source 10.

[0034] V₂: denotes a voltage at the positive terminal 15 of the second energy source 14.

[0035] I₂: denotes a current at the positive terminal 15 of the second energy source 14.

[0036] P₂: denotes a power output from the second energy source 14.

[0037] V₂₀: denotes an open circuit voltage of V₂(i.e., V₂when I₂=0).

[0038] R₂: denotes an equivalent internal resistance of the second energy source 14.

[0039] V_D1: denotes a voltage across the first diode 32.

[0040] V_D2: denotes a voltage across the second diode 33.

[0041] V_L1: denotes V₁minus V_D1.

[0042] V_L2: denotes V₂minus V_D2.

[0043] V_L10: denotes an open-circuit value of V_L1(i.e., V_L1when I₁=0).

[0044] V_L20: denotes an open-circuit value of V_L2(i.e., V_L2when I₂=0).

[0045] V_L: denotes a voltage across the load 30 (also a voltage at the node 61).

[0046] I_L: denotes a current through load 30.

[0047] P_L: denotes a power delivered to the load 30.

[0048] From the preceding definitions:

[0049] V_L1=V₁−V_D1  (1)

[0050] V_L2=V₂−V_D2  (2)

[0051] I_L=I₁+I₂  (3)

[0052] P_L=I_LV_L  (4)

[0053] Note that V_L1is the voltage V₁at the positive terminal 11 of the first energy source 10, decremented by the voltage V_D1across the first diode 32. Similarly, V_L2is the voltage V₂at the positive terminal 15 of the second energy source 14, decremented by the voltage V_D2across the second diode 33.

[0054] V₁and I₁are related by a steady-state characteristic function V₁(I₁). Similarly, V₂and I₂are related by a steady-state function equation V₂(I₂). V_D1and V_D2are typically of the order of about a volt and are generally small in comparison with V₁and V₂, respectively. In principle, both V_D1and V_D2have steady-state characteristic functions, V_D1(I₁) and V_D2(I₂) respectively, which could be taken into account for computational purposes. Using the characteristic functions V₁(I₁) and V_D1(I₁), a characteristic function V_L1(I₁) could be inferred from Equation (1). Using the characteristic functions V₂(I₂) and V_D2(I₂), a characteristic function V_L2(I₂) could be inferred from Equation (2).

[0055] The power management system 5 is designed such that at low load power P_L, or equivalently low load current I_L, all of the load current I_Lis supplied by the first energy source 10, provided that the first energy source 10 is able to supply the total load current I_Lto the load 30. When the first energy source 10 is unable to meet load demand, then the second energy source 14 supplies to the load 30 whatever portion of I_Lcannot be supplied by the first energy source 10. FIGS. 6-9 illustrate the relative contributions of first energy source 10 and the second energy source 14 in supplying the load current I_Lto the load 30 under steady-state conditions (FIGS. 6-8) and transient conditions (FIG. 9).

[0056] FIG. 6 illustrates V_L1versus I₁under steady-state conditions. The V_L1versus I₁curve depends on the internal characteristics of the first energy source 10 and the internal characteristics of the first diode 32. Note that V_L1and I₁are in an inverse relationship. In particular V_L1and I₁are related according to:

[0057] V_L1=V_L10−R₁I₁  (5)

[0058] wherein R₁is the equivalent internal resistance of the first energy source 10. Thus, the curve V_L1versus I₁is linear if R₁is constant, and is non-linear if R₁has a dependence on I₁.

[0059] FIG. 7 illustrates V_L2versus I₂. The V_L2versus I₂curve depends on the internal characteristics of the second energy source 14 and the internal characteristics of the second diode 33. Note that V_L2is nearly flat and thus not very sensitive to I₂, because the equivalent internal resistance R₂of the second energy source 14 is assumed to be small. In particular V_L2and I₂are related according to:

[0060] V_L2=V_L20−R₂I₂  (6)

[0061] Equation (6) shows that R₂is the slope of the V_L2versus I₂curve, and thus explains why V_L2is nearly flat if R₂is sufficiently small. Indeed, V_L2approaches a constant value as R₂approaches 0. Accordingly, V_L2can maintain a voltage close to its open circuit value of V_L20substantially independent of I₂, and thus substantially independent of load power P_L. The V_L2versus I₂curve is valid under both steady state and transient conditions, because it is assumed herein that the second energy source 14 has a capability of responding instantaneously to a sudden increase in power demand.

[0062] It is assumed that the open circuit voltage V₁₀exceeds the open circuit voltage V₂₀, and also that the open circuit voltage V_L10exceeds the open circuit voltage V_L20. Thus at very low load current I_L, I₁>0 and I₂=0. As the load current I_Lincreases, I₂=0 will continue to be satisfied so long as V_L1exceeds V_L2. Generally, under both steady state and transient conditions, I₂=0 if V_L1>V_L2. Based on V_L1versus I₁in FIG. 6, however, V_L1decreases as I₁increases. When V_L1decreases to an extent that V_L1=V_L20, the current in second circuit element 8 begins to flow (i.e., I₂>0).

[0063] FIG. 8 illustrates the load 30 voltage V_Lversus the load current I_L. The V_Lversus I_Lcurve of FIG. 8 combines the V_L1versus I₁curve of FIG. 6 with the V_L2versus I₂curve of FIG. 7, using Equation (3) and assuming V_L10>V_L20. A transition point 70 in the V_Lversus I_Lcurve is determined by the condition V_L1=V_L20. The curve segment 71 to the left of the point 70 is a portion of the V_L1versus I₁curve of FIG. 6 such that V_L1>V_L20. The load current at the point 70 is labeled as I_L0. I_L0is determined by setting V_L1=V_L20and I_L=I_L0in Equation (5) and solving for I_L0to obtain:

[0064] I_L0=(V_L10−V_L20)/R₁  (7)

[0065] If P_L0denotes the load power corresponding to the point 70, then P_L0is related to I_L0in accordance with P_L0=I_L0V_L20. The curve segment 72 to the right of the point 70 is the V_Lversus I curve such that I₁>0 and I₂>0, and may be derived by setting V_L1=V_Lin Equation (5), setting V_L2=V_Lin Equation (6), and substituting Equations (5) and (6) into Equation (3) to yield:

[0066] V_L=I₀+R₀I_L  (8)

[0067] wherein I₀=(V_L10/R₁+V_L20/R₂)R₀, and wherein R₀=1/(1/R₁+1/R₂). If V_L20<V_L1(e.g., point 74), then the first energy source 10 supplies the load current I_L(i.e., I_L74) to the load 30. If V_L20≧V_L1(e.g., point 75), then the load current I_L75=I₁+I₂is supplied to the load 30 by the first energy source 10 and the second energy source 14 in accordance with Equation (5) and Equation (6), respectively; i.e.,

[0068] I₁=(V_L10−V_L75)/R₁  (9)

[0069] I₂=(V_L20−V_L75)/R₂  (10)

[0070] I₁may be expressed as I_L0+ΔI₁. Accordingly, ΔI₁/I₂may be computed by subtracting I_L0in Equation (7) from I₁in Equation (9), and then dividing by I₂in Equation (10) to yield:

[0071] ΔI₁/I₂=R₂/R₁  (11)

[0072] Equation (11) shows that if R₂=0 then ΔI₁=0, I₁=I_LO, and I₂=I_L75−I_L0. If R₂/R₁<<1, which may often occur, then ΔI₁is a small, but finite, component of I₁.

[0073] FIG. 9 illustrates FIG. 8 for a change in power demand and explains how power demand is satisfied under both steady state and transient conditions. If the power management system 5 is initially operating at the point 74 under steady state conditions, then V_L20<V_L1is satisfied and the first energy source 10 supplies the required load current I_L(i.e., I_L74) to the load 30. Assume that under an increase in load power, the new steady state operating point is point 79. If the change in load power from point 74 to point 79 takes place continuously and sufficiently slowly that the first energy source 10 is able to continuously keep up with (i.e., supply energy, or supply the fuel that generates the energy) the increase in power demand, then the transition from point 74 to point 79 proceeds along the curve segment 71, and during the entire transition, V_L20<V_L1is satisfied such that all of the required load current I_Lis continuously supplied by the first energy source 10.

[0074] Alternatively in FIG. 9, if the change in load power from point 74 to point 79 takes place abruptly or rapidly such that the first energy source 10 is unable to keep up with the increase in power demand, then the transition from point 74 to point 79 does not proceed along the curve segment 71 but instead proceeds along a path that is below the curve segment 71, such as the path 80. If a portion of the path 80 is below the V_L20level as shown, then at each point on the path 80 that is at or below the V_L20level, the second energy source 14 will supply part of the load current I_Lto the load 30 such that the first energy source 10 and the second energy source 14 collectively supply the required load current. For example, the point 78 on the path 80 is below the V_L20level. Thus at the point 78, the first energy source 10 supplies the load current I_L78to the load 30, and the second energy source 14 supplies the load current (I_L79−I_L78) to the load 30. As a result, the total load current supplied to the load 30 is I_L79(i.e., I_{L78 +(IL79−IL78})). Thus the first energy source 10 supplies a fraction F of the required load current I_L79wherein F=I_L78/I_L79. The second energy source 14 supplies a fraction (1-F) of the required load current I_L79. Note that at each point on path 80 above the V_L20level, load current is supplied only by the first energy source 10, and the first energy source 10 does not supply all of the required load current I_L79to the load 30, since the first energy source 10 is unable to keep up with the increase in power demand. For example, the point 83 on path 80 is above the V_L20level. Thus, the second energy source 14 does not supply any current to the load 30 at the point 83. Accordingly, the first energy source 10 supplies a current I_L83to the load 30, wherein I_L83is less than the required load current I_L79. In practice, operation on path 80 at the point 83 above the V_L20level is over a time span that represents a lag time for triggering current production by the second energy source 14 when V_L1on path 80 is lowered to below the V_L20level.

[0075] In summary, if V_L20<V_L1is satisfied, then all of the required load current I_Lis supplied by the first energy source 10, in both steady state and transient operational modes. Additionally if V_L20≧V_L1(or V_L2≧V_L1), then the first energy source 10 supplies a fraction F of the required load current I_Lto the load 30, and the second energy source 14 supplies a remaining fraction (1-F) of the required load current to the load 30. In steady state or during a slow transient in which the first energy source 10 is able to keep up with change in load demand, F=1 if the first energy source 10 is at an operating current level less than I_L0, and F=I_L0/I_Lif the first energy source 10 is at an operating current level less of at least I_L0. During transient operation in which the first energy source 10 is unable to keep up with change in load demand, F is calculated as discussed supra in conjunction with the points 78 and 83 of path 80.

[0076] The open circuit voltage V₂₀of the second energy source 14 may be set to a value that depends on a power variable such as a first reference power level P_LREF, wherein P_LREFis a function of a load power representation P_L(t), and wherein t denotes time. Examples of how P_LREFmay be obtained from knowledge of P_L(t) will be given infra. P_L(t) may be in any form and therefore may be an analytical function, a tabular function, a graphical function, etc. P_L(t) may be obtained in a variety of ways. For example P_L(t) may include actual load power history data of the intended operating environment or a similar operating environment. As another example, P_L(t) may include a projection in time based on planned energy requirements of the intended operating environment.

[0077] Once P_LREFis specified, V_L20and I_L0may be determined by simultaneously solving the power equation P_LREF=I_L0V_L0and Equation (7), such as by simultaneously solving said power equation and Equation (7) analytically, numerically, graphically, etc. Then V₂₀can be calculated via V₂₀=V_L20+V_D2(see Equation (2)).

[0078] Any useful criterion may be used for determining a value of P_LREF. A first useful criterion is setting P_LREFequal to P_LAVEwherein P_LAVEis a time average of P_L(t) subject to a weighting function W(t); i.e.,

[0079] P_LAVE=∫W(t)P_L(t)dt/∫W(t)dt  (12)

[0080] over a suitable time domain. If P_L(t) is defined for discrete vales of time, such as in a table, then the integrations in Equation (12) are replaced by summations over the discrete values of time in a manner understood by one of ordinary skill in the art. If P_L(t) is defined over a continuous time interval (e.g., t_A≦t≦t_B), then the suitable time domain for Equation (12) may be any subset of the continuous time interval (e.g., t_C≦t≦t_Dsuch that t_C≧t_Aand t_D≦t_B). If P_L(t) is defined for discrete values of time (e.g., t₁, t₂. . . , t_N), then the suitable time domain for Equation (12) may be any subset of the discrete values of time (e.g., t_J, t_J+1, . . . , t_K, wherein J is any of 1, 2, . . . , and N−1, and wherein K is any of J+1, J+2, . . . , N).

[0081] A choice of W(t) is W(t)=1 which reduces P_LAVEto an unweighted time average of P_L(t). W(t) may also chosen as a function of dP_L(t)/dt (e.g., W(t)=(dP_L(t)/dt)^Zwhere Z is a real positive number such as 0.5, 1, 2, etc.) in consideration of the fact that times t at which the power P_L(t) changes may be more important for determining V₂₀than are times at which P_L(t) is invariant or changes slowly. If W(t)=(dP_L(t)/dt)^Zis used, then it may be desirable to filter out noise and erratic data from P_L(t) before using Equation (12), since the noise may produce non-physical values of dP_L(t)/dt. Additionally, P_L(t) may be preprocessed to filter out preselected high peak powers if it should be decided that the second energy source 14 should always generate positive current I₂during such preselected high peak powers. Such preprocess filtering would avoid an undesired bias in P_LAVE.

[0082] As another illustration of W(t), in a workplace (e.g., a factory), P_L(t) may be expected to have two distinct components: P_L1(t) during working hours and P_L2(t) outside of working hours. While P_L1(t) and P_L2(t) may each have a time dependence, P_L1(t)>>P_L2(t) is expected and P_LREFmay be calculated by Equation (12) by setting W(t)=1 during working hours and W(t)=0 outside of working hours, in recognition of the fact that power requirements outside of working hours may be extraordinarily low and should therefore not be allowed to bias P_LAVE.

[0083] As an alternative to setting P_LREFequal to P_LAVE, one may calculate a standard deviation P_LSDof P_L(t) subject to the weighting function W(t); i.e.,

[0084] P_LSD=[∫W(t){P_L(t)}²dt/∫W(t)dt]^½  (13)

[0085] over the same suitable time domain as is used for Equation (12). If P_L(t) is defined for discrete values of time, such as in a table, then the integrations in Equation (13) are replaced by a summations over the discrete values of time in a manner understood by one of ordinary skill.

[0086] Using Equations (12) and (13) for P_LAVEand P_LSDrespectively, P_LREFis set in accordance with the following formula:

[0087] P_LREF=P_LAVE+ZP_LSD  (14)

[0088] wherein Z is a real number such that Z≧0. If Z is about equal to 0 then Equation (14) simplifies to P_LREFis about equal to P_LAVE. Z affords control of a probability of P_Lexceeding P_LREF. For example, using cumulative normal probability distribution tables, if Z=1 then P_Lis expected to exceed P_LREFabout 15.9% of the time, whereas if Z=2 then P_Lis expected to exceed P_LREFabout 2.3% of the time, and whereas if Z=3 then P_Lis expected to exceed P_LREFabout 0.1% of the time.

[0089] In general, P_LREFmay be selected in any manner that exploits the known or estimated power P_L(t) and does not have to be a function of P_LAVE. For example, P_LREFmay be selected as a fraction B of a peak of P_L(t), wherein B may be within a range of about, inter alia, 0.5 to 1.0. The preceding examples illustrate a few of the may different ways that P_LREFcan be chosen for determining V₂₀such that V₂₀is a function of the load power P_L(t).

[0090] The preceding discussion concerned setting V₂₀to a value that depends on P_LREF, wherein P_LREFis a function of the load power representation P_L(t). Alternatively, V₂₀may be set to a value that depends on a power variable P_1REF, wherein P_1REFis a function of the power output (P₁) of the first energy source 10. Examples of how P_1REFmay be obtained from knowledge of P₁will be given infra.

[0091] Once P_1RFFis specified, a corresponding power reference level P_1REFassociated with P_L1may be inferred from P_L1REF=P_1REF−P_D1. Then V₂₀and I_L0may be determined by simultaneously solving the power equation P_L1REF=I_L0V_L0and Equation (7) which expresses I_L0, such as by simultaneously solving said power equation and Equation (7) analytically, numerically, graphically, etc. Then V₂₀can be calculated via V₂₀=V_L20+V_D2(see Equation (2)).

[0092] Any useful criterion may be used for selecting a value of P_1REFand many choices for P_1REFare possible. A permissible choice is, inter alia, P_1REF=HP_1R, wherein P_1Ris a rated power (or a maximum utilizable power) of the first energy source 10, and wherein H is a fraction such as, inter alia, in a range of between about 0.25 and about 1. H=1 corresponds to choosing P_1REF=P_1Rwhich sets V₂₀to essentially the fuel cell voltage at the rated power of the fuel cell. H may be chosen close to 1 (e.g. 0.90<H<0.95) so as to conservatively avoid exercising the limiting capability of the first energy source 10. If efficiency E(P₁) is a known function of power, then P_1REFcan be set to a value of P₁that maximizes the efficiency E(P₁). There are numerous methods of measuring efficiency E(P₁) such as, inter alia, minimizing cost, maximizing profit, etc. If the first energy source 10 includes fuel cells and the second energy source 14 includes batteries, then it is generally more economical to have peak power loads supplied by the batteries rather than by the fuel cells. There may be constraints added as supplementary conditions to be satisfied such as, inter alia, placing an upper limit on a rate of fuel depletion for situations in which fuel availability is limited. The preceding examples illustrate a few of the may different ways that P_LREFcan be chosen for determining V₂₀such that V₂₀is a function of the power output (P₁) of the first energy source 10.

[0093] As explained supra, V₂₀may be set to a value that depends on P_LREFor P_1REF. Additionally, V₂₀may be set to a value that depends on a combination of P_LREFand P_1REF, such as, inter alia, the linear combination αP_LREF+βP_1REF, wherein α and β are each constant. A numerical value of V₂₀is case dependent. In a fuel cell V₂₀may have a value between, inter alia, 50 volts and 60 volts in some applications (e.g., in some, but not all, residential energy applications of fuel cells), but may have other voltage values in other applications.

[0094] FIG. 2 illustrates FIG. 1 with the second energy source 14 represented as a battery 20, and after a battery charger 24 has been coupled to the battery 20 and is in parallel with the battery 20. The battery charger 24 in FIG. 2 is receiving its input power from the first energy source 10 via circuit lines 51 with a DC to DC converter 26 providing a voltage to the battery charger 24 at a voltage level that is needed to charge the battery 20. The first energy source 10 is able to provide power for charging battery 20 only at such times when first energy source 10 has available reserve power; i.e., when the load power P_Lis less than the rated power P_1Rof the first energy source 10 decremented by a power supplied to the first diode 32 (i.e., P_L<P_1R−I₁V_D1) Alternatively, the battery charger 24 could be an independent battery charger that is not coupled to the first energy source 10, wherein the circuit lines 51 and the DC to DC converter 26 would be removed from the power management system 5 in FIG. 2.

[0095] FIG. 3 illustrates FIG. 2 after the second diode 33 has been replaced by a silicon controlled rectifier (SCR) 34. A purpose of the SCR 34 is to be able to disconnect the battery 20 when the battery charge falls below a given percentage (e.g., about 10% to about 20%) of a maximum battery charge, or equivalently, when the battery voltage V₂falls below a given battery voltage (i.e., the given percentage of maximum battery charge corresponds to the given battery voltage). For example, assume that initially the battery 20 is fully charged at V₂₀of 54 volts and the gate of the SCR 34 is receiving an activation signal (e.g., a positive voltage, a train of positive current pulses, etc.) from an external device (not shown), and that the battery 20 is to be disconnected at a battery voltage of 48 volts or less. When the load 30 becomes sufficiently high that I_L≧I_LO, the SCR 34 becomes forward biased and the SCR 34 is turned on in light of the activation signal still being applied to the gate of the SCR 34. At this point, the battery 20 generates a positive current (i.e., I₂>0). As the battery 20 discharges, the battery voltage V₂will be reduced in accordance with the battery discharging, and the SCR 34 will remain turned on until the SCR 34 becomes reversed biased. With the circuit configuration of FIG. 4, the SCR 34 will become reversed biased when the load 30 becomes sufficiently low that I_L<I_L0(or equivalently, the voltage at node 61 falls below the battery voltage V₂minus the SCR 34 voltage drop). Thus, a deactivation signal (e.g., a negative voltage) should be applied from the external device to the gate of the SCR when the battery voltage V₂falls to 48 volts and lower, so that the SCR 34 is turned off, and remains off, when the SCR 34 becomes reversed biased. When the SCR 34 is so turned off, the battery 20 can be recharged by the battery charger 24 via a charging current from the battery charger 24 into the positive terminal 21 of the battery 20. While the battery is charging, the SCR 34 cannot be inadvertently turned on even if the load 30 is increased such that the SCR 34 becomes forward biased, because of lack of an activation signal at the gate of the SCR 34. When the battery 20 becomes maximally charged at 54 volts, an activation signal from the external device is applied to the gate of the SCR, and a full battery discharge-charge cycle has been completed.

[0096] As just described, a turning off of the SCR 34 at a given battery voltage (e.g., 48 volts) cannot be totally controlled because a reverse biasing the SCR 34 (as the battery 20 discharges) depends on the magnitude of the load 30. The SCR 34 can be turned off at the given battery voltage, however, if a bias circuit element is coupled to the anode and cathode of the SCR 34 such that the bias circuit element has voltage source of a negative polarity which is capable of reverse biasing the SCR 34. The bias circuit element could be turned on (e.,g., via a switch) to reverse bias the SCR 34 when the battery voltage falls to 48 volts, which turns off the SCR 34. In the preceding manner, the battery voltage can be kept within the 48-54 volt range. Thus, the bias circuit element and the SCR 34 collectively maintain the battery charge between a given percentage of maximum battery charge and the maximum battery charge, or equivalently, maintain the battery voltage V₂between a given battery voltage and the maximum battery voltage V₂₀. Generally, the battery voltage may be kept within a desired range from about V₂₀−ΔV to about V₂₀. A useful value of ΔV may be, inter alia, between about 4 volts and about 10 volts.

[0097] FIG. 4 illustrates FIG. 3 after a DC to DC converter 36 has been added in series with the battery 20. An output voltage V_OUTof the DC to DC converter 36 can be preset to a predetermined value, or dynamically controlled by a signal processor as will be illustrated infra in FIG. 5. Returning to FIG. 4, the DC to DC converter 36 may be used to limit a maximum current I_1MAXfrom the first energy source 10. Referring to FIG. 8, I_1MAXis I_L0(corresponding to the point 70) if the DC to DC converter 36 is absent. If it is desired to lower I_1MAXto I_L74of FIG. 8, then V_OUTshould be set to V_L74of FIG. 8. In that manner, the battery 20 will output current to the load 30 whenever the load voltage V_Lis at least V_L74, which (as shown in FIG. 8) constrains the current from the first energy source 10 to less than I_L74. In a similar fashion based on FIG. 8, V_OUTcould be lowered from above V_L74to a value below V_L74to increase I_1MAXto a value greater than I_L74.

[0098] An important feature of FIG. 4 is that the DC to DC converter 36 is not in series with the first energy source 10, but in series with the battery 20. Thus the DC to DC converter 36 draws current (and thus dissipates power) only during the limited time periods during which the battery 20 is furnishing current to the load 30. Hence, power dissipation associated with current flow through the DC to DC converter 36 does not impose a significant energy penalty in comparison with an overall energy delivered to the load 30. If the DC to DC converter 36 were instead in series with the first energy source 10, then the DC to DC converter 36 would draw current at all times, since the first energy source 10 furnishes current to the load 30 at all times. Accordingly, the present invention utilizes the DC to DC converter 36 in an efficient manner that substantially reduces the power losses in DC to DC converter 36, as compared with using the DC to DC converter 36 in series with the first energy source 10. It should be noted that there is no resistive element (aside from the first diode 32) in series with the first energy source 10. Accordingly, during operation of the power management system 5 when the battery 20 is not generating the current I₂, there are no power losses needed to control the current from the first energy source 10, which further improves an efficiency of the power management system 5.

[0099] An additional feature of the DC to DC converter 36 relates to the fact that output voltage V_OUTof the DC to DC converter 36, rather than the output voltage V₂of the battery 20, becomes the voltage against which the output V_L1associated with the first energy source 10 is compared for determining whether the battery 20 will output the battery current I₂. This additional feature enables the output voltage V₂of the battery 20 to have any desired magnitude instead of being required to be less than V_L1.

[0100] FIG. 5 illustrates FIG. 4 after addition of a signal processor 40 for transmitting a signal to the DC to DC converter to dynamically control an output current of the first energy source 10. As discussed supra in conjunction with FIG. 4, an output voltage V_OUTof the DC to DC converter 36 can be raised to decrease the maximum current I_1MAXfrom the first energy source 10, or (conversely) be lowered to increase I_1MAXThe signal processor 40 sends a signal to the DC to DC converter 36 that causes V_OUTto be set to a particular value. The signal processor 40 may include, inter alia, hardware and software that dynamically computes V_OUTin accordance with an algorithm that does not require input data, or in accordance with input data into the signal processor 40 wherein said data is utilized by the algorithm. The algorithm may be hardware controlled or software controlled, or both. FIG. 5 illustrates the situation in which the algorithm requires input data, by showing a particular instrumental structure that could be used to feed the input data into the signal processor 40. In particular, FIG. 5 shows a current transducer 38 which measures the current I₁that flows from the first energy source 10. Let the measured value of I₁be denoted as I_1MEAS. I_1MEASand a setpoint current I_SETPOINTare each fed into a comparator 42 which compares I_1MEASwith I_SETPOINTand sends a result of the comparison to the signal processor 40. The signal processor 40 then acts upon the result of the comparison, in accordance with its algorithm, to determine a new value of V_OUT. Then the signal processor 40 sends a signal to the DC to DC converter 36 which causes V_OUTto be set to the new value. The following two examples illustrate usage of the instrumental structure of FIG. 5. In the first example, it is desired to maintain I₁at a constant value of 200 amp. If I_1MEASis 205 amps then V_OUTshould be raised by an amount that decreases I₁to 200 amp, but if I_1MEASis 195 amps then V_OUTshould be lowered by an amount that increases I₁to 200 amp. In the second example, it is desired to maintain I₁at a maximum value of 200 amp. If I_1MEASexceeds 200 amps then V_OUTshould be raised by an amount that decreases I₁to 200 amp, but if I_1MEASis less than or equal to 200 amps then V_OUTshould not be changed.

[0101] Note that the instrumental structure of FIG. 5 for dynamically supplying input data to the signal processor 40 is merely illustrative, and many other instrumental structures could be utilized for dynamically supplying input data to the signal processor 40.

[0102] It should be noted that while FIGS. 4 and 5 depicted the second energy source 14 of FIG. 1 as the battery 20, the battery 20 in FIGS. 4 and 5 could be replaced by any second energy source 14 referred to, or contemplated, in the discussion supra of FIG. 1.

[0103] It should be noted that while FIGS. 1-5 and the associated text illustrated and described the power management system 5 as having a first energy source 10 and a second energy source 14 (or the battery 20 as representing an embodiment of the second energy source 14), the system power management 5 may generally include the first energy source 10 in parallel with N secondary energy sources, wherein N≧1, wherein if N>1 then the N secondary energy sources are in parallel with one another, and wherein each secondary energy source is of a same type as the second energy source 14 (or the battery 20) of FIGS. 1-5. All features described herein in relation to the second energy source 14 (or the battery 20) are applicable to each secondary energy source of the N secondary energy sources.

[0104] As an example of secondary energy sources with N=2, consider a restaurant in Saratoga Springs, N.Y. in consideration of power requirements as a periodic function of time with a cycle time of one year. Here, hours during which the restaurant is closed will not be taken into account in planning a power management system, since power requirements are relatively very low when the restaurant is closed. Consider three periods of a year: a first period of September-May (lowest average power), a second period of June-July (intermediate average power), and a third period of August (highest average power). Average power for the second period will be higher than for the first period, because the second period occurs during hot weather when air conditioning is required. Average power for the third period will be higher than for the second period, because of hot weather and a substantial increase in restaurant activity due to an high influx of tourists who visit Saratoga Springs for horse racing at the Satatoga Racetrack which occurs only during the month of August. Here, a first secondary energy source could be a first battery and a second secondary energy source could be a second battery, wherein the second battery is in parallel with the first battery, and wherein the first battery is in parallel with the first energy source 10. An open-circuit voltage V_L20(1) associated with the first battery could be set at a level such that the first battery is substantially activated (i.e., generates a current outflow) during the second period of June-July (and not during the first period except for infrequent high load power excursions) and supplements the first energy source 10. An open-circuit voltage V_L20(2) associated with the second battery could be set at a level higher than V_L20(1) such that the second battery is substantially activated (i.e., generates a current outflow) during the third period of August and supplements both the first energy source 10 and the first battery. The preceding example is merely illustrative. There are numerous applications for which two or more secondary energy source may be useful in the context of the present invention.

[0105] While particular embodiments of the present invention have been described herein for purposes of illustration, many modifications and changes will become apparent to those skilled in the art. Accordingly, the appended claims are intended to encompass all such modifications and changes as fall within the true spirit and scope of this invention.