Renewable energy power generation systems

Abstract

We describe a modular adjustable power factor renewable energy inverter system. The system comprises a plurality of inverter modules having a switched capacitor across its ac power output, a power measurement system coupled to a communication interface, and a power factor controller to control switching of the capacitor. A system controller receives power data from each inverter module, sums the net level of ac power from each inverter, determines a number of said capacitors to switch based on the sum, and sends control data to an appropriate number of the inverter modules to switch the determined number of capacitors into/out of said parallel connection across their respective ac power outputs.

Claims

1. A distributed energy generation inverter system for applying volt-ampere reactive (VAR) control to an alternating current (AC) output, the distributed energy generation inverter system comprising: a plurality of individual inverters, each individual inverter comprising a direct current (DC) power input for receiving DC power from a DC source, an AC power output for supplying AC power to a common AC circuit, and a reactive element switchable to be connected in parallel to the AC output; and a controller communicatively coupled to each of the plurality of inverters, the controller configured to monitor the individual power output of each inverter, determine a fraction of the maximum total potential power being supplied by the plurality of inverters, and when the fraction is above a pre-determined threshold, selectively controlling one or more of the individual inverters to switch their respective reactive elements into parallel connection across the respective AC outputs.

2. The distributed energy generation inverter system of claim 1, wherein the controller employs phase control to selectively control the one or more of the individual inverters to switch the respective reactive elements in parallel for a portion of an AC cycle in which the reactive element is switched in.

3. The distributed energy generation inverter system of claim 1, wherein the common AC circuit is a three phase AC power feed, and wherein an inverter of the plurality of inverters is for coupling to each phase of the three phase AC power feed, the controller configured to control a power factor of each phase of the three phase AC power feed.

4. The distributed energy generation inverter system of claim 1, wherein the reactive element includes a capacitor.

5. The distributed energy generation inverter system of claim 1, wherein the controller is communicatively coupled to each of the plurality of inverters by way of a wireless communication channel.

6. The distributed energy generation inverter system of claim 1, wherein the DC power source is a solar photovoltaic panel.

7. The distributed energy generation inverter system of claim 1, wherein the pre-determined threshold is greater than or equal to 0.5.

8. The distributed energy generation inverter system of claim 1, wherein the controller is configured to access a lookup table that stores power factor compensation data defining a number of reactive elements to switch in parallel to the fraction of the maximum total potential power being supplied by the plurality of inverters.

9. The distributed energy generation inverter system of claim 1, wherein monitoring the individual power output of each inverter includes sensing a current provided by each of the inverters to the common AC circuit.

10. The distributed energy generation inverter system of claim 1, wherein each of the individual inverters includes a switching device coupled to switch their respective reactive element, and wherein the controller is configured to switch the switching device at a peak voltage point of the AC power.

11. The distributed energy generation inverter system of claim 9, wherein each of the switching devices includes a triac.

12. The distributed energy generation inverter system of claim 1, wherein the controller includes a field-programmable gate array (FPGA).

13. The distributed energy generation inverter system of claim 1, wherein each of the individual inverters includes a two-stage power converter, wherein a first stage of the two stage power converter is electrically isolated from the second stage of the two stage power converter.

14. The distributed energy generation inverter system of claim 13, wherein the first stage is to be coupled to the DC power source, and wherein the first stage include a switching DC-to-AC converter.

15. The distributed energy generation inverter system of claim 1, wherein each of the plurality of individual inverters includes a second reactive element switchable to be connected in parallel to the AC output, and when the fraction is above a pre-determined threshold, the controller is further configured to selectively control one or more of the individual inverters to switch their respective second reactive elements into parallel connection across the respective AC outputs.

16. An controller for a distributed energy generation system for applying volt-ampere reactive (VAR) control to an alternating current (AC) output, the controller comprising: a communication interface for communicating with a plurality of inverters; processing logic coupled to the communication interface; and a computer-readable medium accessible to the processing logic, the computer-readable medium storing instructions, when executed by the processing logic will cause the controller to perform a method comprising: monitoring, via the communication interface, the individual power output of each of the plurality of inverters; determining a fraction of the maximum total potential power being supplied to a common AC power grid by the plurality of inverters; selectively switching, via the communication interface, reactive elements in one or more of the plurality of inverters when the fraction is above a pre-determined threshold, wherein selectively switching the reactive elements puts the reactive elements in parallel connection with the AC power grid to apply volt-ampere reactive (VAR) control to an AC output to the AC power grid.

17. The controller of claim 16, wherein selectively switching the reactive elements includes switching the reactive elements for a portion of an AC cycle to facilitate phase control.

18. The controller of claim 16, wherein the communication interface is a wireless communication interface.

19. The controller of claim 16, wherein the pre-determined threshold is greater than or equal to 0.5.

20. The controller of claim 16, wherein the computer-readable medium includes a lookup table that stores power factor compensation data defining a number of reactive elements to switch in parallel to the fraction of the maximum total potential power being supplied by the plurality of inverters.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) These and other aspects of the invention will now be further aspects of the invention will now further described, by way of example only, with reference to the accompanying figures in which:

(2) FIGS. 1a and 1b show, respectively, power factor requirements according to German standard VDE4105, and an outline block diagram of an example power conditioning unit;

(3) FIGS. 2a and 2b show details of a power conditioning unit of the type shown in FIG. 1b;

(4) FIGS. 3a and 3b show details of a further example of solar photovoltaic inverter in which an input power converter incorporates an LLC resonant power converter;

(5) FIGS. 4a to 4c show, respectively, single phase and three phase modular adjustable power factor solar converter systems according to embodiments of the invention, and an example illustration of a residential property fitted with the system;

(6) FIGS. 5a-5c show, respectively, block diagrams showing increasing detail of a solar inverter module for use with the system of FIG. 4;

(7) FIG. 6 shows a flow diagram of a procedure for implementation by a system controller of the system of the system of FIG. 4; and

(8) FIGS. 7a and 7b show, respectively, a graph of system displacement power factor against the target requirement of FIG. 1a, and the power factor error in tracking the curve of FIG. 1a.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(9) VAR control is required where grid requirements mandate leading or lagging control over the output of any grid connected power source. Typical requirements provide limits in the range.

(10) FIG. 1a shows VDE4105 requirements in which the power factor of systems above 3.68 kVA must be continuously controlled based on percentage of full load system power. The power factor for systems above 13.8 kVA follows the 0.9 curve, those below 13.8 kVA follow the 0.95 curve.

(11) Further, power sources must be capable of being programmed to a specific power factor in overexcited and under-excited conditions with a latency of 10 seconds. We address this by employing feedback from the sub-source, a micro-inverter, in the context of a combined system of multiple micro-inverters' connected to grid. In embodiments the ability to close this feedback loop and to control power factor (per phase in a three phase system) to within 0.01 (cos φ) enables high accuracy and control at the grid connection point.

Power Conditioning Units

(12) By way of background and context, to assist in understanding the operation of embodiments of the invention we first describe an example photovoltaic power conditioning unit. Thus in FIG. 1b shows photovoltaic power conditioning unit of the type we described in WO2007/080429. The power converter 1 is made of three major elements: a power converter stage A, 3, a reservoir (dc link) capacitor C.sub.dc 4, and a power converter stage B, 5. The apparatus has an input connected to a direct current (dc) power source 2, such as a solar or photovoltaic panel array (which may comprise one or more dc sources connected in series and/or in parallel). The apparatus also has an output to the grid main electricity supply 6 so that the energy extracted from the dc source is transferred into the supply. Capacitor C.sub.dc is preferably non-electrolytic, for example a film capacitor.

(13) 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 C.sub.dc 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.

(14) 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 C.sub.dc into the mains. The circuit topology may be either a voltage source inverter or a current source inverter.

(15) FIG. 2 shows details of an example of a power conditioning unit of the type shown in FIG. 1b; like elements are indicated by like reference numerals. In FIG. 2a Q1-Q4, D1-D4 and the transformer form a dc-to-dc conversion stage, here a voltage amplifier. In alternative arrangements only two transistors may be used; and/or a centre-tapped transformer with two back-to-back diodes may be used as the bridge circuit.

(16) In the dc-to-ac converter stage, Q9, D5, D6 and Lout perform current shaping. In alternative arrangements this function may be located in a connection between the bridge circuit and the dc link capacitor: D.sub.6 acts as a free-wheeling diode and D.sub.5 prevents current form flowing back into the dc-link. When transistor Q.sub.9 is switched on, a current builds up through L.sub.out. When Q.sub.9 is switched off, this current cannot return to zero immediately so D.sub.6 provides an alternative path for current to flow from the negative supply rail (D.sub.5 prevents a current flowing back into the dc-link via the body diode in Q.sub.9 when Q.sub.9 is switched off). Current injection into the grid is controlled using Q.sub.9: when Q.sub.9 is turned on the current flowing through L.sub.out increases and decreases when it is turned off (as long as the dc-link voltage is maintained higher than the grid voltage magnitude). Hence the current is forced to follow a rectified sinusoid which is in turn unfolded by the full-bridge output (transistors Q.sub.5 to Q.sub.8). Information from an output current sensor is used to feedback the instantaneous current value to a control circuit: The inductor current, i.sub.out, is compared to a reference current, i.sub.ref, to determine whether or not to switch on transistor Q.sub.9. If the reference current is higher than i.sub.out then the transistor is turned on; it is switched off otherwise. The reference current, i.sub.ref, may be generated from a rectified sinusoidal template in synchronism with the ac mains (grid) voltage.

(17) Transistors Q5-Q8 constitutes an “unfolding” stage. Thus these transistors Q5-Q8 form a full-bridge that switches at line frequency using an analogue circuit synchronised with the grid voltage. Transistors Q5 and Q8 are on during the positive half cycle of the grid voltage and Q6 and Q7 are on during the negative half cycle of the grid voltage.

(18) Thus in embodiments the power conditioning unit comprises a generic dc-ac-dc that provides voltage amplification of the source to above the grid voltage, and isolation, and a current source inverter (CSI) connected to the mains. The current injection is regulated using current shaping (current-control) in the inductor of the CSI via the intermediate buck-type stage. (This is described further in our GB2415841B, incorporated by reference).

(19) Control (block) A of FIG. 1b may be connected to the control connections (e.g. gates or bases) of transistors in power converter stage A to control the transfer of power from the dc energy source. The input of this stage is connected to the dc energy source and the output of this stage is connected to the dc link capacitor. This capacitor stores energy from the dc energy source for delivery to the mains supply. Control (block) A may be configured to draw such that the unit draws substantially constant power from the dc energy source regardless of the dc link voltage V.sub.dc on C.sub.dc.

(20) 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 V.sub.dc on C.sub.dc.

(21) The capacitor C.sub.dc 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 C.sub.dc. 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).

(22) 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 FIG. 2b. In this example these blocks operate independently but share a common microcontroller for simplicity.

(23) 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.

(24) In a photovoltaic power conditioning unit the microcontroller of FIG. 2b will generally implement an algorithm for some form of maximum power point tracking. In embodiments of the invention we describe later this or a similar microcontroller may be further configured to control whether one or both of the dc-to-dc power converter stages are operational, and to implement “soft” switching off of one of these stages when required. The microcontroller and/or associated hardware may also be configured to interleave the power transistor switching, preferable to reduce ripple as previously mentioned.

(25) Now referring to FIG. 3a, this shows a further example of a power conditioning unit 600. In the architecture of FIG. 3 a photovoltaic module 602 provides a dc power source for dc-to-dc power conversion stage 604, in this example each comprising an LLC resonant converter. Thus power conversion stage 604 comprises a dc-to-ac (switching) converter stage 606 to convert dc from module 602 to ac for a transformer 608. The secondary side of transformer 608 is coupled to a rectifying circuit 610, which in turn provides a dc output to a series-coupled output inductor 612. Output inductor 612 is coupled to a dc link 614 of the power conditioning unit, to which is also coupled a dc link capacitor 616. A dc-to-ac converter 618 has a dc input from a dc link and provides an ac output 620, for example to an ac grid mains supply.

(26) A microcontroller 622 provides switching control signals to dc-to-ac converter 606, to rectifying circuit 610 (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. (The skilled person will be aware of many ways in which such sensing may be performed). In some embodiments the microcontroller 622 implements a control strategy as previously described. As illustrated, the microcontroller 622 is coupled to an RF transceiver 624 such as a ZigBee™ transceiver, which is provided with an antenna 626 for monitoring and control of the power conditioning unit 600.

(27) Referring now to FIG. 3b, this shows details of a portion of an example implementation of the arrangement of FIG. 3a. This example arrangement employs a modification of the circuit of FIG. 2a and like elements to those of FIG. 2a are indicated by like reference numerals; likewise like elements to those of FIG. 3a are indicated by like reference numerals. In the arrangement of FIG. 3b an LLC converter is employed (by contrast with FIG. 2a), using a pair of resonant capacitors C1, C3.

(28) The circuits of FIGS. 1 to 3 are particularly useful for 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. 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.

VAR Control Techniques

(29) We will now describe embodiments of a modular adjustable power factor solar inverter system which is able to track the power factor curve of FIG. 1a and, more particularly, which is able to provide a power factor which is adjustable in the range plus/minus 0.95 cos φ with, in embodiments, an accuracy of 0.01 cos φ. A power conditioning unit (solar inverter) of the type described above, with a controllable current source (Q9) and a series output inductance in the above example) followed by a grid frequency unfolding rectifier is able to provide a lagging power factor which can approach unity. This therefore addresses the −0.95 cos φ requirement. Broadly speaking in embodiments of the system a leading power factor is provided by employing a switched capacitor arrangement (with control of the switching phase), but there are some additional techniques which are employed to facilitate control and upgrade of a modular system.

(30) Thus referring to FIG. 4a, this shows an embodiment of a solar inverter system 400 comprising a set of PV panels 402 each providing dc power to a respective solar inverter or power conditioning unit 404, preferred embodiments of the type described above. Each inverter provides an ac grid mains output to a shared ac connection 406 which provides a common grid power feed 408. In embodiments of the system there are multiple solar inverters, each with a relatively small power output, as described further later. The system is controlled by system controller 410, for example using a Zigbee™ wireless network coupling an antenna 412 of the system controller to each of the inverters.

(31) Each inverter includes a switched capacitor coupled to a power factor controller under control of system controller 410. As previously described, each inverter controls the output current and reads the RMS output voltage and thus is able to determine the percentage of its full power that the inverter is providing and/or an absolute measure of the power it is providing to the common grid tie 406, 408. Thus system controller 410 is able to determine the absolute power provided to the grid by each inverter and/or the percentage of an inverter's full power being provided by each inverter. The system controller 410 uses this information to control the switching of the capacitors in one or more of the inverters, as described later.

(32) FIG. 4b illustrates a three phase solar inverter system 420 similar to that of 4a, in which like elements are indicated by like referencing numerals. In the example of FIG. 4b one set of inverters 404a, b, c, powered by respective sets of solar PV panels 402a, b, c, drives each phase of the grid mains, and each set of inverters driving each phase is separately controlled by system controller 410.

(33) In embodiments of the three phase system shown in FIG. 4b optionally inverter modules 404 may provide information to the system controller 410 which enables the system controller to determine which inverter modules are connected to which individual phase of the three phase grid supply. This may comprise, for example, data identifying the timing of the of the sinusoidal voltage waveform on the ac output of the inverter, this information being sent to system controller 410 over the wireless network. From this timing information the system controller is able to determine which inverters are connected to which phase and therefore is able to determine which sets of inverters are to be controlled together to control the displacement power factor for each phase. Alternatively, information defining which inverter is connected to which phase may be programmed into the system controller, for example on installation.

(34) FIG. 4c illustrates an example installation of a system of the type shown in FIGS. 4a and 4b (in which only one solar panel/inverter is shown for simplicity). In the arrangement of FIG. 4c the function of the system controller 410 is implemented in a gateway 430, here a wireless base station, which also provides an interface to the solar inverter system. In this example the gateway 430 is coupled to a broadband internet modem/router 432 which provides a connection to the Internet 434. This in turn provides a remote interface 436 for a monitoring site, for example for a user or system vendor or electricity supplier.

(35) The arrangement of FIG. 4c shows the panels on the roof 442 of a building 440 with an electricity cable 406 to a consumer unit configured to provide the grid interface 408. The illustrated example Zigbee™ repeater 444 with one antenna 446 external to the building, and one antenna 448 internal to the building, together with an optional signal splitter 450 or other arrangement. This facilitates an RF signal propagation between the system controller and the inverters, as described in more detail in our co-pending U.S. patent application Ser. No. 13/244,222 (hereby incorporated by reference), which is useful where the roof incorporates conductive thermal insulation or, in a commercial building, comprises corrugated metal sheets.

(36) Referring now to FIG. 5 this shows a portion of an example solar inverter module 404 for the system of FIG. 4, in which like elements to those previously described are indicated by like reference numerals. FIG. 5 shows just the output portion of the solar inverter previously described, that is the portion of the inverter from the dc link onwards. Because FIGS. 5a to 5c show a similar solar inverter modules with an increasing level of detail, for simplicity these will be described together.

(37) Thus the inverter 404 comprises a dc link 614 with an energy storage capacitor 616 which provides power to a current source stage 500, more particularly a voltage controlled current source providing power to a buck inductor assembly 612, and thence to a full wave rectifier output unfolding stage 618. In embodiments, as previously described, the current injection is regulated using current control in the inductor assembly 612 via an intermediate buck-type stage provided by current source 500 (this circuit block also includes a microcontroller and unfolding drivers, not explicitly illustrated for simplicity). The inductor (output) current is sensed by resistor 504 and compared with the reference to determine whether or not to provide current to inductor 612, thus providing current mode control. Further details can be found in our U.S. patent application Ser. No. 11/718,879, hereby incorporated by reference.

(38) The ac output is filtered by capacitor 506 and inductor 508 and protected by varistor 510 and fuse 512 prior to ac grid mains output 514. For simplicity details of the drivers for unfolding stage 618 are shown as part of circuit block 500, using an ac phase sense connection 516 to synchronise with the grid.

(39) Continuing to refer to FIG. 5, circuit block 500 also includes an output 520 for controlling whether or not one or more capacitors 522 are to be switched into parallel connection across ac output 514. The switched capacitor drive output 520 may, in embodiments, be provided as an output from the microcontroller 622 illustrated in FIG. 3, receiving a signal from the system controller 410 via RF transceiver 624. The switched capacitor drive output 520 is provided to a switched capacitor control circuit block 530 via an isolation transformer 524, the control block 530 responding to the switch signal to connect or disconnect capacitor 522 across ac output 514. in embodiments the switched capacitor control block 530 may implement on/off control of switching of the capacitor 522, or phase-sensitive control based on detection of 0-crossing of the capacitor current (peak detection of the grid voltage).

(40) FIG. 5c shows details of an embodiment of the arrangement using a triac 540 to control switching of capacitor 522. Use of a triac is preferable as this is better able to handle a surge than a MOSFET. In this example implementation a transistor stage 542 couples an output of transformer 524 to a triac driver integrated circuit 544, for example an AVS08 from SGS-Thomson or similar. In embodiments this is arranged to provide a train of pulses to turn the triac on at the peak of the grid voltage (0 capacitor current), although in principle because the triac will hold itself on only one pulse need be employed if this is timed to turn the triac on when the current through the triac is rising.

(41) The circuit of FIG. 5c also able to provide variable phase control for the switched capacitor drive output 520. This control enables the proportion of an ac cycle for which the capacitor 522 is connected across the ac mains to be varied, drive 520 thus determining a proportion of a cycle for which the capacitor is applied. This is achieved by controlling a switching point of the triac 540. in this way the effective capacitance provided across the ac grid output of the inverter module may be varied, in embodiments linearly.

(42) In one preferred method of triac control the triac is first switched in at the zero-crossing of the grid voltage. Then, for several grid cycles, for example of order 50 cycles, the triac is driven with a continuous pulse train to assure firing at all angles while the switching current transient reduces. After the switching transient has reduced to negligible levels, the triac may be driven either with a single pulse or a pulse train only near the peak of the grid voltage.

(43) As previously described, in preferred embodiments the inverter has the ability to sense or otherwise determine the phase of the grid voltage. Although this is not essential it is helpful, in particular for controlling triac switching. Sensing the phase of the grid voltage can help to assure adequate firing of the triac during the startup transient. This reduces stress on the VAR-control capacitor, by facilitating switching in at the zero-crossing. It also reduces power loss—by reducing the triac drive to the zero-crossing of the current, which corresponds to the peak and trough of the grid voltage after the initial current transient has been reduced to negligible levels.

(44) The example circuits of FIG. 5 merely illustrate one preferred embodiment of a solar inverter module for use in the system we describe. The skilled person will appreciate that alternative approaches may also be employed and that, in particular, embodiments of the technique we describe for power factor control are not restricted to use with a solar inverter employing a current source followed by an inductor assembly.

(45) Referring next to FIG. 6, this shows an embodiment of a procedure operating on system controller 410 which controls the solar inverter modules to switch the output capacitors in or out so that the system is able to provide a power factor which varies from unity to plus 0.95 cos φ leading. In broad terms, capacitance of 0, 1 or more modules are switched into the grid as required based on the output export power of the system, as this varies from a minimum to a maximum. Because the system comprises multiple micro-inverters connected in parallel feeding a common grid tie point (per phase in the case of a three phase system) the required capacitance is also distributed amongst the modules and if, for example, the system is extended by adding more modules, the ability to correct the power factor is also, automatically, increased.

(46) Thus referring to FIG. 6, after initialisation (S650) the system controller polls each inverter to read the net power output from the inverter via the Zigbee™ network (S652). The controller then sums the power reported by each inverter module (S656) and calculates the power being provided by the system as a percentage of the total output power. The power read from each inverter may either be an absolute power or a percentage of the total power that particular inverter is able to provide; in this latter case the percentages may be averaged to determine the total output percentage of the system. As previously mentioned, the power factor of an inverter module (not including the capacitor) is substantially unity, but not precisely unity. Thus there may be an optional step (S654) to adjust for this non-unity power factor by means of a lookup table. This may either be performed at the inverter or in the system controller.

(47) Once the total output power of the system as a percentage of the maximum export power is determined, the system can then lookup or otherwise determine the required system power factor (S658), for example to approximate the curve of FIG. 1a. The system then determines the number of inverters on which to switch in the output capacitance, to achieve this (S660). In one embodiment each inverter module has the same output capacitance, provided as a single switched output capacitance, but in alternative approaches an inverter may provide two or more switchable capacitances and/or different inverter modules may have differently sized switchable capacitances. The skilled person will appreciate that account may be taken of these variations in step 660. The system controller then sends capacitor switch command signals to a corresponding number of inverters (S662), and the procedure loops back to step S652.

(48) Where phase control of the output control capacitor is provided by an inverter module step S660 may also determine a proportion of the cycle for which the capacitor is applied, this information being transmitted to the inverter modules at step S662.

(49) FIG. 7a shows examples of the results of the procedure of FIG. 6 for a solar inverter system of the type illustrated in FIG. 4. The staircase-form curve 700 illustrates the effect of on/off capacitor control and the linear curve 702 illustrates the effect of phase (proportional) control of the switched capacitors of the inverter modules. FIG. 7b illustrates the error in displacement power factor (cos φ) for the staircase curve of FIG. 7a, showing the variation of this error with overall system power exported.

(50) As can be seen from FIG. 7 and the preceding description, the power factor is controlled in steps (or linearly) by switching capacitors inside each inverter at the appropriate system power level. In embodiments of the system the net displacement power factor can be controlled in theory to within an arbitrary tolerance limit by using the power output of each inverter to be small enough to ensure an adequate number of steps. Thus for example for a system to provide a net displacement power factor within the 1% of a required value each step should cause the displacement power factor to change by at most 2% (thus changing from the target value−1% to the target value+1%). Preferably this step size should be a little smaller than this limit, to accommodate tolerances and hysteresis in the system.

(51) In the example given above, to satisfy VDE4105 with a 3.68 kVA system the (maximum) inverter size is 736 Watts, leading to a system with 5 such inverters. In practice it is preferable to use inverters with slightly lower power and/or to provide each inverter with two or more capacitor steps, to allow some margin for error. Thus in this example a practical upper limit in inverter size for each inverter is approximately 480 Watts for inverters that each employ just a single switched VAR control capacitance.

(52) Continuing the above example, for systems above 13.8 kVA in overall output power level, the VAR compensation required is twice that at 3.6 kVA to 13.8 kVA (a compensation of cos φ up to 0.10 rather than 0.05). The inverters can be arranged to provide this level of VAR compensation either by providing two separately switchable VAR control capacitances in each inverter module and/or by using smaller inverters in the system, for example in the range 240 Watts-300 Watts.

(53) Broadly speaking we have described a solar inverter system in which one or more switched VAR control capacitances are provided in each solar inverter module, sized according to the maximum output power of the inverter module, to achieve a desired power factor compensation target. These are combined together in a system with a system controller which is able to remotely control the addition of this capacitance and/or the phase of the capacitance, thus controlling the overall power factor of the system. Thus, for example, embodiments of this technique are able to control the displacement power factor of the overall system with an accuracy of 0.01 cos φ, or better. Embodiments of the system are closed loop in the sense that they monitor inverter power output into the common grid connection and provide control data back to the inverter modules for controlling the addition/phase of the switched capacitance, but are open loop in the sense that, in embodiments, no measurement is needed of the power factor at the grid connection. The techniques we have described are applicable to both single phase and three phase architectures.

(54) No doubt many other effective alternatives will occur to the skilled person. 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.