This invention relates to power conditioning units for delivering power from a dc source to an ac output, either for connecting directly to the mains (grid) utility supply or for powering mains (grid) devices directly, independent from the mains utility supply. The techniques we describe are particularly suitable for power conditioning units for photovoltaic (PV) modules.
We have previously described a range of improved techniques for increasing reliability and efficiency in photovoltaic power conditioning units (inverters), for example in WO2007/080429 and in many other of our published patent applications.
We now describe some further techniques for increasing the efficiency of harvesting power from a PV panel, in particular at low levels of power output. These in turn facilitate harvesting of solar energy early and late in the day and in overcast conditions, thus increasing the overall energy yield
According to the present invention there is therefore provided a power conditioning unit for delivering power from a dc power source to an ac mains power supply output, the power conditioning unit comprising: a dc power input to receive dc power from a dc power source; an ac power output for delivering ac power to said ac mains power supply; a dc link coupled between said dc power input and said ac power output; at least a first dc-to-dc converter coupled between said dc power input and said dc link; and a dc-to-ac converter coupled between said dc link and said ac power output; wherein said dc-to-dc converter comprises: a transformer having an input winding and an output winding; an input dc-to-ac converter coupled between said dc power input and said input winding of said transformer; an ac-to-dc converter coupled between said output winding of said transformer and said dc link; and wherein said output winding of said transformer has a winding tap between first and second portions of said output winding, and wherein said ac-to-dc converter comprises: first and second rectifiers, each connected to a respective first and second said portion of said output winding, to said dc link, and to said winding tap; and a series inductor connected to said winding tap; and wherein said rectifiers are each connected to said winding tap of said output winding via said series inductor such that said series inductor is shared between said first and second rectifiers.
Broadly speaking in embodiments by arranging for a shared inductor, component tolerances can be more closely controlled, thus reducing the need to oversize portions of the circuitry to take account of component value variations. Moreover, in embodiments the shared series inductor facilitates the use of two or more dc-to-dc converters on the input side of the power conditioning unit by, in effect, forcing current sharing between these converters (by limiting dips in the current from each), which provides further efficiencies described later.
In some preferred embodiments the shared inductor comprises a winding on a core of the transformer, which provides further advantages both by reducing losses and also because the inductance is effectively reflected back to primary side of the transformer so that errors in the resonant frequency of the dc-to-ac converter, again due to component tolerances (of the primary side inductance) are effectively swamped and dominated by the output side shared inductor. This again facilitates accurate control of the operation of the inverter, hence further increasing efficiency. More particularly, the output side inductance is reflected back, in proportion to the square of the primary:secondary terms ratio and appears in series with the primary side inductance, the combined inductance defining the resonant frequency of operation (which is generally fixed by design). Thus embodiments of the techniques we describe facilitate accurate setting of this resonant frequency, and hence efficient operation.
In one preferred implementation the inductor is built into the transformer by winding the inductor on a portion of the core of the transformer which shares a magnetic circuit with both the first and second portions of the output winding. More particularly the transformer may have three legs, one bearing a first portion of the output winding, a second bearing a second portion of the output winding, and a third, for example central, leg bearing the winding of the shared inductor. In embodiments a portion of this shared inductor winding is also wound around each of the first and second legs in an opposite sense to that of the output windings of the transformer, to cancel the respective portions of the output winding flux.
In some particularly preferred embodiments the power conditioning unit includes a pair (or more) of the dc-to-dc converters operating in parallel between the dc power input and the dc link. Providing two such converters reduces the I2R losses in the power conditioning unit. In embodiments each of these has a dc-to-ac input converter comprising one or more power switching devices driving one or more input windings of the transformer. Each of these dc-to-dc converters also comprises a coupling/isolation transformer. Optionally the two (or more) dc-to-dc converters may each have a rectifying circuit (comprising one or more rectifiers) following the transformer, or they may share a common rectifying circuit or circuits.
In embodiments each of these power switching devices may free-run, but preferably switching of these devices is synchronised, in particular to reduce ripple on the dc input capacitor(s) and hence smooth the overall load on the dc input. In principle operating the switching devices of the two converters 180° out of phase will reduce the ripple on the input capacitor by a factor of 4. For example this is possible if, in effect, switching devices of the input side dc-to-dc converter and those of the output side dc-to-ac converter are separate and do not need to be synchronised. However in embodiments the input and output rectifiers are synchronised, in which case it can be preferable to operate the switching devices of the input dc-to-ac converters 90° out of phase with one another (as between the two converters), which achieves a factor of 2 reduction in ripple on the input capacitance.
In some particularly preferred embodiments a dc conversion stage controller is provided to control operation of one or both of the dc-to-dc converters, in particular in response to the sensed input or output power. In this way the efficiency of the power conditioning unit can be optimised taking into account both I2R losses and core losses in the transformers. The transformer core loss is typically substantially constant, fixed by the applied volt seconds that is by the voltage and frequency. Thus at low power levels switching one converter off can reduce this fixed component of loss (typically perhaps of order 1-10 Watts), and at larger powers running both converters to reduce the I2R losses. In embodiments the controller is configured to switch off one dc-to-dc converter if, when doing so, the overall losses would be substantially the same or reduced.
In preferred embodiments the dc conversion stage controller is configured to provide a ‘soft’ switching of a dc-to-dc converter so that stored energy in the transformer of the converter has time to dissipate gradually rather than stressing the circuit components. This may be achieved, for example, by gradually changing the frequency of operation of a converter to move the frequency off resonance and/or by gradually reducing the duty cycle of a PWM (pulse with modulation) switching control waveform to increase the switching device off time.
In principle sensing the power available from the PV panel could be performed using a series resistor, but this would incur losses. In embodiments, therefore, an analogue of the actually available power is measured, in particular by measuring a level of ripple (at the resonant frequency) on a capacitor of an input dc-to-ac converter (more particularly a capacitor in series with one of the power switching devices. With such an arrangement less ripple indicates less available power and vice versa.
In embodiments each of the dc-to-dc converters is substantially matched to the other, although this is not essential. For example, one converter could have a high power rating, the other a lower power rating in order to skew the efficiency curve towards the high or low power output end. In practice the determination of a threshold at which to switch operation from just a single converter to a pair of converters is best determined by experiment, for example plotting efficiency curves for the one-converter and two-converter modes to determine where the optimum changeover point is located. The skilled person will recognise that a range of different variations of converter circuits may be employed, using full bridge or half bridge rectifier configurations, synchronous rectifiers or diodes, and so forth. Thus whilst in some preferred embodiments the first and second rectifiers comprise synchronous rectifiers, in particular MOSFETs, in other embodiments the first and second rectifiers may simply each comprise a diode; and alternatively the rectifiers may comprise a more complex rectifier circuit.
In a related aspect the invention provides a photovoltaic power conditioning unit for delivering power from a dc power source to an ac mains power supply output, the photovoltaic power conditioning unit comprising: a dc power input to receive dc power from a photovoltaic power source; an ac power output for delivering ac power to said ac mains power supply; a dc link coupled between said dc power input and said ac power output; a dc-to-ac converter coupled between said dc link and said ac power output; first and second dc-to-dc power converters, each coupled between said dc power input and said dc link and each having a series inductor in a connection between a said dc-to-dc power converter and said dc link; and a dc conversion stage controller, coupled to sense an input or output power of said power conditioning unit and to control said first and second dc-to-dc converters; and wherein said dc conversion stage controller is configured to control, responsive to said sensed power, whether one or both of said dc-to-dc converters operates to feed dc power from said dc input to said dc link.
As described above, use of the series inductor helps to equalise current sharing between the first and second dc-to-dc power converters. The dc link will typically comprise a positive and negative (or ground) roll and a series inductor may be included in series with either or both of a power supply connection and a return connection between the dc link and a dc-to-dc power converter. As previously mentioned, in some particularly preferred embodiments the series inductor is implemented as a winding on a transformer of the dc-to-dc converter, in particular connected to a tap on the output winding of the transformer. Again as previously described, preferably the controller is configured to implement ‘soft’ switching of a converter to reduce component stresses; and again in embodiments the operation of switching devices in the converters is interleaved, preferably at 90° or 180° phase difference between the two converters.
In a further related aspect the invention provides a method of operating a power conditioning unit, the method comprising: converting dc power from a dc power source to dc power for a dc link of said power conditioning unit using a pair of dc-to-dc power converters by converting said dc power to ac, providing said ac to a transformer and converting ac from said transformer to dc for said dc link: converting dc power from said dc link to ac output power; coupling dc outputs of said dc-to-dc power converters via one or more inductors to force load current sharing between said dc-to-dc power converters; sensing a power level into or out of said power conditioning unit; and inhibiting operation of a said dc-to-dc power converter responsive to said sensed power level to reduce losses in a core of said transformer of said dc-to-dc power converter below a threshold power level, and above said threshold power level reducing losses due to resistive heating in a said dc-to-dc power converter by enabling operation of both said dc-to-dc power converters.
Preferably the inductor is built into a transformer of the dc-to-dc power converter.
The invention further provides a power conditioning unit comprising means for performing the steps of the above described method.
These and other aspects of the invention will now be further described, by way of example only, with reference to the accompanying figures in which:
Power Conditioning Units
By way of background, helpful for understanding the operation of embodiments of the invention, we first describe an example photovoltaic power conditioning unit. Thus
The power converter stage A may be, for example, a step-down converter, a step-up converter, or it may both amplify and attenuate the input voltage. In addition, it generally provides electrical isolation by means of a transformer or a coupled inductor. In general the electrical conditioning of the input voltage should be such that the voltage across the dc link capacitor Cdc is always higher than the grid voltage. In general this block contains one or more transistors, inductors, and capacitors. The transistor(s) may be driven by a pulse width modulation (PWM) generator. The PWM signal(s) have variable duty cycle, that is, the ON time is variable with respect to the period of the signal. This variation of the duty cycle effectively controls the amount of power transferred across the power converter stage A.
The power converter stage B injects current into the electricity supply and the topology of this stage generally utilises some means to control the current flowing from the capacitor Cdc into the mains. The circuit topology may be either a voltage source inverter or a current source inverter.
Control (block) A of
Control (block) B may be connected to the control connections of transistors in the power converter stage B to control the transfer of power to the mains supply. The input of this stage is connected to the dc link capacitor and the output of this stage is connected to the mains supply. Control B may be configured to inject a substantially sinusoidal current into the mains supply regardless of the dc link voltage Vdc on Cdc.
The capacitor Cdc acts as an energy buffer from the input to the output. Energy is supplied into the capacitor via the power stage A at the same time that energy is extracted from the capacitor via the power stage B. The system provides a control method that balances the average energy transfer and allows a voltage fluctuation, resulting from the injection of ac power into the mains, superimposed onto the average dc voltage of the capacitor Cdc. The frequency of the oscillation can be either 100 Hz or 120 Hz depending on the line voltage frequency (50 Hz or 60 Hz respectively).
Two control blocks control the system: control block A controls the power stage A, and control block B power stage B. An example implementation of control blocks A and B is shown in
In broad terms, control block A senses the dc input voltage (and/or current) and provides a PWM waveform to control the transistors of power stage A to control the power transferred across this power stage. Control block B senses the output current (and voltage) and controls the transistors of power stage B to control the power transferred to the mains. Many different control strategies are possible. For example details of one preferred strategy reference may be made to our earlier filed WO2007/080429 (which senses the (ripple) voltage on the dc link)—but the embodiments of the invention we describe later do not rely on use of any particular control strategy.
In a photovoltaic power conditioning unit the microcontroller of
An example configuration of transformer 306 is shown in the inset diagram, as illustrated comprising a magnetic core including a gap for increased reluctance/energy storage. The leakage inductance of transformer 306 is shown explicitly in
The circuit of
Improved Techniques
The circuit of
Referring now to
In
As described in more detail later, in a power conditioning unit the dc output of the converter of
Referring next to
Broadly speaking, the value of the output inductance when integrated into transformer 356 can be broadly the same as when a separate output inductor is employed (although parasitic leakage inductance may have a second order effect, changing the resonant frequency of the switching power converter circuit). Broadly speaking the value of the output inductance relates to the power rating of the inverter, and should also support the desired output current. For example in a ‘microinverter’ with an output power in the range <500 Watts to 1000 Watts the output inductor of each of a pair of front end inverter switching converter circuits may be of order 200 μH.
Broadly speaking, when designing the transformer for a particular power conditioning unit, a useful starting point is the primary side volt-seconds. These are determined by the electronic design (dc operating voltage and switching frequency), and the desired output inductance. The volt-seconds in turn allow the centre leg flux to be determined (noting that the ampere-turns in the centre leg are twice those in the outer legs, because half the flux goes each way). The centre leg flux then allows the effective area of the core and, more particularly, the core material and dimensions, to be determined.
Referring to
Using this notation, the magnetizing flux is given by:
The inductive flux is given by:
And thus:
φ1,2=φL1,2−φm1,2
And the centre leg flux, is:
For an ideal transformer:
The magnetizing currents are:
im1=−im2
And the primary currents are:
ipri1,2=ip1,2+im1,2
ipri1=−ipri2
Now referring to
φm1+φm2=0
φL1+φL2=φC
From
From
From
And from
We will now describe some preferred embodiments of photovoltaic power conditioning units which employ multiple front end power converter stages connected in parallel between the dc input of the power conditioning unit and the dc link of the power conditioning unit. In embodiments each of these power converter stages implements a boost/isolation stage using an LLC conversion topology. One motivation for this paralleled approach is to mitigate the dc-to-ac conversion losses between the photovoltaic power source (which may be either a string-connected or individual PV module) and the grid. A significant source of such power losses is in the initial boost/isolation stage of such a converter, that is between the dc input and dc link of the power conditioning unit. The losses in this stage relate to both ac losses, and dc losses, for example I2R losses.
We will now describe techniques to mitigate such losses, in embodiments using phase offset switching of the converter stages to minimise ac ripple current losses as well as dc losses. The techniques we describe enable increased efficiency, as well as reduced component stress and higher reliability. More particularly, in embodiments using such a paralleled resonant/flyback/hard-switched power topology reduces the dc current per converter stage, thus reducing dc losses, and multiphase switching reduces RMS (root mean square) ripple, both reducing aggregate power loss.
In embodiments of the techniques we describe each front end converter stage may operate substantially independently in the sense that it may be separately enabled or disabled, to thereby shape the efficiency curve of the combined system. This is particularly beneficial in a system which is operating at less than 100% of its available power (rating) we will also describe techniques for soft switching of a power converter during enable/disable of a converter, for example by frequency modulation and/or duty cycle modulation.
In embodiments each converter stage is designed to operate at 50% of the maximum load, plus some additional margin to accommodate the accuracy of sharing between stages, including component tolerance and drift over temperature. However in preferred embodiments we use output inductance to in effect force 50/50 sharing between the two power converter stages, which reduces the effects of temperature and component drift and can produce a more efficient and cost effective system. The skilled person will appreciate, however, that 50/50 sharing is not essential.
Thus referring now to
A microcontroller 622 provides switching control signals to dc-to-ac converters 606a, b, to rectifying circuits 610a, b (for synchronous rectifiers), and to dc-to-ac converter 618 in the output ‘unfolding’ stage. As illustrated microcontroller 622 also senses the output voltage/current to the grid, the input voltage/current from the PV module 602, and, in embodiments, the dc link voltage. In some preferred embodiments the microcontroller 622 implements a control strategy as previously described, although the operation of embodiments of the invention is not tied to any particular control strategy or, for example, to any particular MPPT (maximum power point tracking) strategy.
In the circuit of
The techniques we describe are particularly useful for so-called microinverters, for example having a maximum rate of power of less than 1000 Watts and or connected to a small number of PV modules for example just one or two such modules. This is because in such systems the panel voltages can be as low as 20 volts and hence the conversion currents can be in excess of 30 amps RMS.
Referring now to
Referring now to
At low input/output powers it is more efficient to run just a single front end converter 604, to reduce core loses in the transformer, but at higher input/output powers it is more efficient to run both converters, to reduce resistive losses. To avoid rapid dumping of the energy stored in a transformer 608 preferably microcontroller 622 is configured to switch a converter on/off gradually, for example either by gradually moving the switching frequency off resonance and/or by gradually reducing the duty cycle of a PWM control signal to the converter switch, to reduce the switch on-time and thus gradually dissipate the stored energy. It will be appreciated that, if desired, such techniques may also be employed to turn a converter partially on or partially off.
The point at which a change-over occurs between running one converter and running both converters is best determined by experiment, for example by plotting curves of efficiency and/or loss when running two converters and when just running a single converter, to determine the change-over point.
Thus
Although this is generally preferable to optimise efficiency, optionally the change-over point may be adjusted away from the maximum efficiency point, for example to take account of external factors and/or to provide some hysteresis in the switching. Optionally the switching point may be biased towards either single or dual converter use (for example in the latter case to reduce overall component stresses and hence potentially prolong lifetime).
We have thus described techniques for potentially maximising solarvoltaeic power conditioning unit efficiency over substantially the full PV voltage range, and additionally techniques which enable incorporation of lossy inductor elements into an integrated magnetic component. Embodiments of the transformer we describe are also relatively straightforward to manufacture.
No doubt many other effective alternatives will occur to the skilled person. For example it is potentially possible to omit the input dc-to-dc converter. It will be understood that the invention is not limited to the described embodiments and encompasses modifications apparent to those skilled in the art lying within the spirit and scope of the claims appended hereto.
Number | Date | Country | Kind |
---|---|---|---|
1104800.6 | Mar 2011 | GB | national |
This Application is a continuation application of U.S. patent application Ser. No. 13/769,257, filed Feb. 15, 2013, now published as U.S. Publication 2013/0229842. U.S. patent application Ser. No. 13/769,257 is a continuation application of U.S. patent application Ser. No. 13/244,161, filed Sep. 23, 2011, now issued as U.S. Pat. No. 8,391,031. U.S. patent application Ser. No. 13/244,161 claims benefit of an earlier-filed United Kingdom Patent Application 1104800.6, filed Mar. 22, 2011. U.S. patent application Ser. No. 13/769,257, now published as U.S. Publication 2013/0229842, U.S. Pat. No. 8,391,031 and United Kingdom Patent Application 1104800.6 are incorporated herein by reference.
Number | Name | Date | Kind |
---|---|---|---|
4183079 | Wachi | Jan 1980 | A |
4488136 | Hansen et al. | Dec 1984 | A |
4772994 | Harada et al. | Sep 1988 | A |
5404059 | Löffler | Apr 1995 | A |
5576941 | Nguyen et al. | Nov 1996 | A |
5636107 | Lu et al. | Jun 1997 | A |
5726615 | Bloom | Mar 1998 | A |
5946206 | Shimizu et al. | Aug 1999 | A |
5982645 | Levran et al. | Nov 1999 | A |
6211657 | Goluszek | Apr 2001 | B1 |
6335871 | Kita et al. | Jan 2002 | B1 |
6445599 | Nguyen | Sep 2002 | B1 |
6807069 | Nieminen et al. | Oct 2004 | B2 |
6888728 | Takagi et al. | May 2005 | B2 |
7471524 | Batarseh et al. | Dec 2008 | B1 |
7518886 | Lai et al. | Apr 2009 | B1 |
7728708 | Fushimi et al. | Jun 2010 | B2 |
8369113 | Rodriguez | Feb 2013 | B2 |
8391031 | Garrity | Mar 2013 | B2 |
8410889 | Garrity et al. | Apr 2013 | B2 |
8743570 | Mueller | Jun 2014 | B2 |
8934269 | Garrity et al. | Jan 2015 | B2 |
20030185026 | Matsuda et al. | Oct 2003 | A1 |
20040076028 | Achleitner et al. | Apr 2004 | A1 |
20040145928 | Takada | Jul 2004 | A1 |
20040189432 | Yan et al. | Sep 2004 | A1 |
20040190314 | Yoshida | Sep 2004 | A1 |
20040233685 | Matsuo et al. | Nov 2004 | A1 |
20070133241 | Mumtaz et al. | Jun 2007 | A1 |
20070164612 | Wendt et al. | Jul 2007 | A1 |
20070217239 | Rottger et al. | Sep 2007 | A1 |
20070290656 | Lee Tai Keung | Dec 2007 | A1 |
20090168461 | Nakahori | Jul 2009 | A1 |
20110007527 | Liu et al. | Jan 2011 | A1 |
20110051468 | Kyono | Mar 2011 | A1 |
20120081937 | Phadke | Apr 2012 | A1 |
Number | Date | Country |
---|---|---|
0340006 | Nov 1989 | EP |
20564382 | May 2009 | EP |
2365599 | Sep 2011 | EP |
2689525 | Jan 2014 | EP |
2419968 | May 2006 | GB |
2486509 | Jan 2013 | GB |
2491494 | May 2013 | GB |
2003289674 | Oct 2003 | JP |
8202134 | Jun 1982 | WO |
0070731 | Nov 2000 | WO |
02078164 | Oct 2002 | WO |
2006089778 | Aug 2006 | WO |
WO 2007080429 | Jul 2007 | WO |
2010119324 | Oct 2010 | WO |
WO 2012127208 | Sep 2012 | WO |
Entry |
---|
English translation, EP2365599, Device for converting electrical energy and method for operating such a device, Mueller Jens-Uwe, pp. 1-12. |
Portions of prosecution history of U.S. Appl. No. 13/244,161, mailed Feb. 1, 2013, Garrity, Paul. |
Portions of prosecution history of U.S. Appl. No. 13/769,257, mailed Aug. 4, 2014, Garrity, Paul. |
Portions of prosecution history of GB 1104800, mailing date Dec. 11, 2012, Enecsys Limited. |
Portions of prosecution history of GB 1211883, mailing date Apr. 9, 2013, Enecsys Limited. |
International Search Report and Written Opinion of PCT/GB2012/050515, mailing date Dec. 14, 2012, Enecsys Limited. |
International Preliminary Report on Patentability of PCT/GB2012/050515, mailing date Oct. 3, 2013, Enecsys Limited. |
Mumtaz, Asim, et al., “Grid Connected PV Inverter using a Commercially Available Power IC,” PV in Europe Conference, Oct. 2002, Rome, Italy. |
Examination and Search Report for related European Application No. 12715705.5, dated Sep. 15, 2016, 14 pages. |
Number | Date | Country | |
---|---|---|---|
20150256092 A1 | Sep 2015 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 13769257 | Feb 2013 | US |
Child | 14537032 | US | |
Parent | 13244161 | Sep 2011 | US |
Child | 13769257 | US |