WO2014071459A1 - Grid stability control system and method - Google Patents

Abstract

A grid stability control system (GSCS) (11) is used in conjunction with a solar power system (SPS) (17), which is a relatively unstable power source, to stabilise power output to an electrical grid (13). The SPS (17) is typically supplementing a diesel power system (DPS) (16) which in combination supply the electrical grid (13). The GSCS (11) comprises a bidirectional inverter system (19), a battery system (21), and a PLC system (23). Control processes at the PLC system (23) provide a deadzone control process to maximise battery open circuit mode duration, a rate of change of output power limiting process to accommodate a grid, an immediate compensation response control mode, and a slow compensation response control mode, and a partial state of discharge targeting process to manage the state of charge of a plurality of battery clusters comprising the battery system.

Description

GRID STABILITY CONTROL SYSTEM AND METHOD

FIELD OF THE INVENTION

This invention relates to a grid stability control system and a method of controlling grid stability being supplied power from a primary power system and being supplemented by a relatively unstable (RU) power source. The invention has particular, although not exclusive, utility in managing a solar power generating plant for supplementing the power supply to a grid from a diesel power generating plant in remote areas.

Throughout the specification, unless the context requires otherwise, the word "comprise" or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers.

BACKGROUND OF THE INVENTION

The following discussion of the background art is intended to facilitate an understanding of the present invention only. It should be appreciated that the discussion is not an acknowledgement or admission that any of the material referred to was part of the common general knowledge as at the priority date of the application.

Some Australian electric utilities operate a substantial number of diesel- fuelled power stations as the primary power system for remote communities that are too distant to connect to the main electricity grid. These communities are often populated by hundreds of residents, and are serviced by local commercial activities, all of which requires civil infrastructure to support the region. The wet season can make it difficult to access some communities by road for up to 1 to 2 months of the year.

With the rising cost of fossil fuel supplies such as diesel, the global community imperative to reduce carbon emissions and Australia's abundant supply of renewable energy, government is requiring that renewable energy sources be used as secondary power systems to supplement the operation of these primary power systems and reduce the amount of fossil fuel used. Photovoltaic (PV) power generation has been identified as the most consistent source of renewable power to cover the daily power consumption profile of these communities.

Incorporating a high level of PV power generation into a diesel power station, known as 'penetration', can:

(i) degrade quality of supply;

(ii) increase churn in multiple diesel genset usage;

(iii) increase spinning reserve margin; and

(iv) result in loss of grid supply.

The principal cause of these disadvantages is when the solar power level changes abruptly, such as from passing cloud on a sunny day. The reason for this is that a power generation system comprising one or more diesel engines requires many seconds, and often minutes, to adjust to substantial power changes on the local grid, depending on the rate of change of power and whether additional generator resource is needed.

There can also be short periods when the level of generated solar power gets close to the grid load consumption level, or where a power distribution feeder fault occurs, which requires the power output of the solar generating plant to be under direct and immediate control of the power station.

Both of these concerns could, if not managed properly, lead to

unacceptable levels of electricity voltage and frequency changes, or possibly even power station outages due to protection devices tripping.

There is accordingly a need for improved systems and methods that at least attempt to address these and other limitations of existing approaches to managing voltage and frequency fluctuations on the described types of electricity networks.

SUMMARY OF THE INVENTION

An object of the present invention is to implement methods of managing the output power of a relatively unstable power source to provide for a more stabilised supply of power. A relatively unstable power source is characterised by having an output that fluctuates with environmental conditions.

The inventive concept arises from a recognition that a system for controlling the stability of a grid that is being supplied power from a relatively unstable power source can be advantageously controlled by operating one or more processes that in isolation or combination seek to at least partially compensate for transient fluctuations detected in the grid.

The present invention in a first aspect provides a controller operating a plurality of processes including a deadzone control process to maximise battery open circuit mode duration and minimise power consumption of the inverters.

The present invention in a second aspect provides a controller operating a plurality of processes including a rate of change of output power limiting process to accommodate a grid.

The present invention in a third aspect provides a controller operating a plurality of processes including:

(i) an immediate compensation response control mode; and

(ii) a slow compensation response control.

The present invention in a fourth aspect provides a controller comprising a battery manager operating a plurality of processes including a partial state of discharge targeting process to manage the SOC of a plurality of battery clusters comprising the battery and inverter system.

BRIEF DESCRIPTION OF DRAWINGS

The best mode for carrying out the invention is described by way of an example of a specific preferred embodiment of the present invention. The description is made with reference to the accompanying drawings.

Fig 1 is a diagrammatical view of a power generation system in accordance with the preferred embodiment.

Fig 2 is a diagrammatical view of the electrical power flow of the power generation system.

Fig 3 is a diagrammatical view of the control signal paths of the power generation system.

Fig 4 is a graphical view of the target control response of the rolling average of the power output of the solar power system.

Fig 5 is a graphical view of the target control response of the power output of the solar power system.

Figs 6A and 6B are block diagrams depicting the various processes and sub-processes performed by the PLC system of a grid stability control system. Figs 7A, 7B and 7C form a set of graphs showing the step disturbance response of output power from the solar power system (SPS) which is described as part of the power response control process.

Fig. 8 is a graph charting how power output from the solar power system is compensated within deadzone limits.

Fig 9 is a flowchart depicting the steps performed by the power response control process.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

A first embodiment of the invention provides a method and system in which a compensation power response is provided when a measured power output of a relatively unstable power system is determined to deviate from recent values by a predetermined negative power value or a predetermined positive power value. System overview

Fig. 1 provides a schematic overview of an electrical grid 13 for supplying power to electrical loads 14. A preferred embodiment of the best mode for carrying out the present invention is directed towards a grid stability control system (GSCS) 11 that forms part of an electrical grid 13.

The two originating sources of electrical power for the electrical grid 13 are a diesel power system (DPS) 16 and a solar power system, (SPS) 17. The SPS 17 comprises photovoltaic panels, and the DPS 16 comprises diesel-powered generator sets.

The SPS 17 is a relatively unstable (RU) power source, as its output fluctuates with environmental conditions, in particular incident sunlight. The DPS 16 forms a mainstay of power generation, as its power output depends only upon its capacity and fuel supply, barring service intervals.

The DPS 16 and SPS 17 are co-ordinated by the GSCS 11. The SPS 17 and GSCS 11 form a site power source 15 that supplements the DPS 16 to supply electrical loads 14. The site power source 15 and DPS 16 in effect form a power station for supplying the electrical load 14.

The GSCS 11 comprises a bidirectional inverter system 19, an energy storage system in the form of a battery system 21 , and a PLC system 23. The inverters used in the bi-directional inverter system 19 are single-phase low voltage grid feeding inverters arranged in groups of three to provide for the delivery of three-phase power from the VRLA battery strings.

The batteries used in the battery system 21 are valve regulated lead acid (VRLA) batteries arranged in multiple strings each comprising a series of connected monoblocks.

The battery strings in the battery system 21 and the inverters in the bidirectional inverter system 19 are grouped into clusters where each cluster contains a battery group comprising multiple battery strings and a group of three inverters.

The PLC system 23 comprises a programmable logic controller (PLC) system with associated communications adaptors to connect to various components of the electrical grid 13 via appropriate communication hardware.

Figs. 2 and 3 are diagrams that relate to the electrical grid 13

schematically depicted in Fig. 1 , insofar as these supplementary diagrams represent the electrical power flow through the electrical grid 13 (Fig. 2), and the control signal paths implemented by the PLC system 23 in the electrical grid 13 (Fig. 3).

Control processes at PLC system

The PLC system 23 includes processes to perform the following principal functions:

• calculate compensation power command levels.

• apportion the compensation power levels to cluster power command levels.

· respond to commands and control levels from the diesel generator power system of the DPS 16.

• control the power limit command level of the SPS 17.

• energise and de-energise the inverters.

• energise and de-energise the battery string strings.

· initiate automated shutdown of processes and devices.

• communicate command levels and read process parameters from the inverters.

• read process parameters from the power meters. • monitor switching device status.

• monitor device fault status.

• monitor battery voltage and current level status.

• energise and de-energise air-conditioners for cooling of the bidirectional inverter system 19.

Figs. 6A and 6B jointly depict the actual processes and sub-processes that are run on the PLC system 23 to perform these functions, set out below as follows.

• Power response control process 31 for calculating compensation power command levels, essentially involving:

o monitoring of SPS power generation level;

o testing for compensation power requirement;

o calculating compensation power level; and o apportioning compensation power level to cluster power levels.

• Cluster power apportioning process 33 for apportioning the compensation power levels to cluster power command levels, essentially involving:

o monitoring cluster power limitation parameters;

o calculating prospective cluster power capability;

o biasing apportioning calculations to reduce cluster churn and maintain maintenance processes; and

o apportioning compensation power level.

• SPS power command process 35 for controlling the power command level to the SPS, essentially involving:

o responding to external control power limit requirements; o calculating SPS power limit level margin;

o communicating SPS power limit level and o monitoring SPS power level.

• Operating mode process 37 for energising and de-energising the inverters and the battery strings, essentially involving:

o activating and deactivating clusters to support peak

prospective compensation power requirement; o activating and deactivating battery strings to support operational capability of each battery string;

o operating battery strings in a partial state of discharge and in open-circuit conditions.

· Remote control processes 39, essentially involving:

o responding to start and stop commands;

o responding to communications watchdog; and o providing operational process levels.

• Battery maintenance process 41 for monitoring battery voltage and current level status, essentially involving:

o monitoring each battery monobloc for non-normal voltage and current conditions,

o automating periodic battery equalisation process,

o automating periodic battery discharge process.

· Facility control process 43 for initiating automated shutdown of processes and devices, and energising and de-energising air- conditioners, essentially involving:

o automated multi-level shutdown responses, o air conditioner control.

Power response control process

The Power response control process 31 monitors the total real rms power level of the SPS 17, Ppv, and compares Ppv to the rolling average real power output Pps of the site power source 15, which corresponds to the site output power.

The GSCS 11 is designed so that a small difference in power level within a compensation deadzone power range does not produce a compensation power response - which is required to minimise the cumulative energy throughput of the battery storage 21 , and is within the normal grid load variation level of the DPS 16.

On the other hand, the GSCS 11 is designed so that a large difference in power level, exceeding the compensation deadzone power range, initiates a compensation power requirement, and a compensation power level is calculated. The compensation power level, when added to the output power of the SPS 17, maintains the real power output of the site power source 15 within immediate power level variation bounds, and longer duration power level ramp rate bounds. The immediate power level variation bounds, and the power level ramp rate bounds, may have different bounds of levels for increasing and decreasing site power source output power.

To manage this process, the PLC system 23 of the GSCS 11 uses a process variable in respect of the site output power, Pps. The rate of change of Pps is selected to be constrained to between (for example) +60kW/minute increasing, and -SR/minute decreasing, where SR is the spinning reserve of the DPS 16.

The power output from the SPS 17, Ppv, is measured by a first AC meter and is typically exported directly from the SPS 17. Any change in Ppv is treated as a main disturbance to the PLC control loop - and change of Ppv is bounded by the maximum prospective power level of the SPS (PRC).

The manipulated variable is the power transferred through the bidirectional inverter 19, Pgsc, and can be:

• positive power flow (that is, with the GSCS 11 acting as a power source, with the battery system 21 discharging), or

· negative power flow (that is, with the GSCS 11 acting as a power sink, with the battery system 21 charging).

Pgsc is mainly comprised of the sum of the power flow through all clusters, Pgsc = Pel + Pc2 + Pc3 etc, where Pci is the power flow through the cluster from the batteries of cluster T. A negative peripheral power transfer also occurs due to auxiliary control functions of the PLC system 23, in particular the air-conditioners. The individual cluster power levels are not required to be the same, and are effectively operated independently.

The maximum prospective power capability of each cluster Pcpi is dependent on the capability of the bi-directional inverter system 19 and the capability of the battery system 21. The sum of Pcpi is compared to the power limit level PRC to determine if the GSCS 11 can cover any future disturbance event (either positive or negative excursion). If the GSCS 11 cannot cover possible events, then the maximum power level of the SPS inverters (as opposed to the bidirectional inverter system 19) is reduced accordingly. This scenario may occur if the state of charge (SOC) of the battery system 21 is too low, or too high, or battery strings are off-line, or clusters are off-line.

Example of large step disturbance of solar power system

Fig. 7 indicates a drop disturbance with a corresponding battery power discharge sawtooth shape, with a leading edge peak power level equal to the drop in Ppv, and with Pgsc ramping at -SR/min. This represents a worst-case battery power excursion to cover a large step disturbance in SPS power Ppv.

The cluster power level is controlled by a real power setpoint with a preset power factor (PF) level.

Power response control process flow

Fig. 9 depicts a flowchart for the Power Response Control process 31 , and involves the following sub-processes, referred to above, and described in further detail below:

• the Test for Active Response Level sub-process 45;

• the Active Response Required sub-process 47;

• the Set Cluster Power Level sub-process 49;

· the SPS Power Limit Setting Required sub-process 51 ; and

• the Set SPS Power Limit Level sub-process 53.

The Test for Active Response Level sub-process 45 is performed as follows. The first AC Meter 20A is read for Total Real Power parameter Ppv (t=0).

The rolling average of Pps (Psra) is calculated using 1 second samples of

Pps as follows: Psra =∑ Pps (t=0 to t=-59) / 60

This uses the present (t=0) and previous 59 samples that make up a 1 minute duration. This is thus a sliding window calculation using a block of measurements taken over the immediately preceding 1 minute period.

The SPS power change ΔΡρν at t=0, is calculated relative to Psra: ΔΡρν =

Ppv - Psra

The Active Response Required sub-process 47 uses the parameters LIMpos and LIMneg to signify the required deadzone limit levels. LIMpos and LIMneg vary with DPS 16, and the present embodiment are selected as LIMpos = 60kW, and LIMneg = 23kW.

A deadzone limit LIMpos, and LIMneg, on either side of Psra is used to alleviate the need to act on small changes in power output of the SPS 17, Ppv - that is, the deadzone limits strive to maintain Pgsc=0 for as often and for as long as practicable.

The sizing of the deadzone limits is a compromise between higher levels of Pps fluctuations, and more frequent periods where Pgsc is non-zero. Non-zero Pgsc accumulates to higher annual levels of cycling of the battery system 21 , and as a consequence power loss.

Natural variations in output of the SPS 17, namely Ppv, are likely to limit the practical minimum levels of LIMneg and LIMpos. A default level of LIMneg may be initially set to 50% of the spinning reserve of DPS 16.

This is achieved by not responding to small levels of ΔΡρν within the deadzone limits, or in other words a deadzone limit band - effectively avoiding minor fluctuations in Ppv from forcing Pgsc from a zero value. If a compensation power response is required by the GSCS 11 , and the Ppv change is too large such that ΔΡρν would exceed the deadzone limit band, when a disruptive change in solar power output Ppv occurs (for example, due to passing cloud cover), then Pps is forced to ramp back to Psra level. As Pps = Ppv + Pgsc, a compensation power response is provided by Pgsc to partly mitigate against the disruptive change to smooth out Pps. Table 1 directly below sets out an example.

TABLE 1

If Pgsc = 0 at t=-1 , and given ΔΡρν = Ppv - Psra at t=0, then:

for ΔΡρν < LIMpos, and for ΔΡρν > - LIMneg, then set ResponseRequired = 0

// that is, compensation power response is not required. for ΔΡρν > LIMpos, and for ΔΡρν < - LIMneg, then set ResponseRequired = 1 // that is, compensation power response is started.

If ResponseRequired = 1 AND ΔΡρν > ZZ ^* LIMpos then Pgsc = - ΔΡρνι + LIMpos/60

If ResponseRequired = 1 AND ΔΡρν < -ZZ ^* LIMneg then Pgsc = - APpvt - LIMneg/60

// that is, a compensation power response is still required. If ResponseRequired = 1 at t=-1 , then at t=0:

for ΔΡρν < ZZ ^* LIMpos OR ΔΡρν > - ZZ ^* LIMneg, then set ResponseRequired = 0

// that is, a compensation power response is stopped.

This test identifies if a compensation power response is required to start (ResponseRequired = 1 ), and identifies when Ppv has returned to a level close to Psra and hence the compensation power response can stop (ResponseRequired = 0). The parameter ZZ is typically selected to be a relatively low value (as an example, a value of for example 1 0% may be suitable, depending upon the site), but not too low that 'jitter' in Ppv keeps Pgsc≠ 0 for longer than needed. A smaller ZZ value ensures that a compensation power response continues to be provided until Ppv is relatively close to Psra, and certainly well within the deadzone band defined as the range between Psra less LIMneg, and Psra plus LIMpos.

A second embodiment of the present invention provides a method and system in which a combined power output - provided by the compensation power response combined with the power output of the relatively unstable power source - is constrained to a maximum predetermined negative rate of change, and a maximum predetermined positive rate of change. The second embodiment of the invention can operate independently of the first embodiment of the invention, or in conjunction with the first embodiment of the invention.

The second embodiment of the invention is substantially the same as the first embodiment of the invention, as described above with reference to the accompanying drawings.

A compensation power response is provided by the GSCS 11 , to compensate when required changes in the power output of the SPS 17. The quantum of the Pgsc that may be provided is denoted in Table 1 above. If a compensation power response is required by the GSCS 11 , but the Ppv change is not large enough to exceed the deadzone band limits, but could exceed the allowable ramp rate for Pps, then Pps is adjusted to ramp at the maximum ramp rate allowable.

The amount of compensation power response Pgsc provides is equal to

ΔΡρν minus half the allowed rate of change, as set forth in the Table 2 directly below.

TABLE 2

RATEpos is rate of power increase over the period of the rolling average for Psra. If ΔΡρν₍ > RATEpos/2 then

Pgsct = - (APpvt - RATEpos/2)

// that is, battery system is charging to absorb a swell in solar power.

RATEneg is rate of power decrease over the period of the rolling average for Psra.

If ΔΡρνκ -RATEneg/2 then

Pgsct = APpvt + RATEneg/2

// that is, battery system is discharging to compensate for a droop in solar power. As such, low-level fluctuations generated by the SPS 17 that are not modified by the GSCS 11 , will be the inherent environmental solar fluctuations reflected in the output of the SPS 17.

A third embodiment of the invention integrates the first embodiment of the invention and the second embodiment of the invention, as described above with reference to the accompanying drawings, and as foreshadowed in relation to the second embodiment of the invention.

Accordingly the third embodiment of the invention provides a method and system in which a compensation power response is provided when a measured power output of a relatively unstable power system is determined to deviate from recent values by a predetermined negative power value or a predetermined positive power value, and a combined power output - provided by the

compensation power response combined with the power output of the relatively unstable power source - is constrained to a maximum predetermined negative rate of change, and a maximum predetermined positive rate of change.

An example of the deadzone band response is shown in a graph depicted as Fig. 8. Ppv exceeds the positive and negative deadzone limits in three separate instances in the minute from 0 to 60s indicated.

Note that the Y-axis represents Ppv, and the origin is positioned at a logged value of Psra (at t=0), and shows that Psra was a lower value 1 minute earlier, and will end up being a higher value 1 minute later. Ppv at the origin has a slightly higher value than Psra in this example. Deadzone limit lines are for convenience drawn for t>0, but rather than remain horizontal (as shown, for convenience of depiction), the deadzone limit lines will drift move up and down as Psra varies for t>0.

When Ppv changes enough to exceed the deadzone limits then compensation power by the GSCS 11 starts, and Pps deviates from Ppv, with Pps ramping back towards Psra until such time as Ppv approximates Pps, when compensation power in effect ceases.

A fourth embodiment of the invention provides a method and system by which management of a battery system is effected when used in conjunction with a bi-directional inverter system and a PLC system for providing a compensation power response to supplement a relatively unstable power source.

Aspects of the fourth embodiment of the invention are touched upon in the preceding description, with reference to the accompanying drawings, in relation to the first, second or third embodiments of the invention.

The fourth embodiment of the invention is preferably provided in a context substantially the same as that of any one of the first, second or third

embodiments of the invention. The fourth embodiment of the invention can integrate any one or more of the first, second or third embodiments of the invention.

Particular additional aspects of the fourth embodiment of the invention are described below with reference to the battery system 21 , which operates in conjunction with a bi-directional inverter 19, under control of a PLC system 23. This forms the GSCS 11 , which supplements the SPS 17, and provides a compensation power response Pgsc as required. The Pgsc is ultimately sources from the battery system 21.

As described above with reference to Fig. 6A, Power response control process 31 involves -broadly - calculating compensation power command levels. This process 31 comprises a Set Cluster Power Level sub- process 49, which uses the following parameters:

• Pcpi - the individual cluster prospective power levels as calculated in the Cluster Capability process 59, which is described in further detail below

• Pci - the individual cluster set power levels, which are communicated to each individual cluster i whenever Pgsc is non-zero

The Set Cluster Power sub-process 49 firstly tests if Pgsc < PF^*Pcp(Max1 ) + Pch(Max2) + Pch(Max3), for a GSCS 11 having three clusters, where

Pcp(Max1 ) is the Pep level of the cluster with the 'worst' SOC (which changes depending on whether the Pgsc is positive or negative) and PF is the power factor setting of that cluster, and Pch(Max2) and Pch(Max3) are the power levels of the other two clusters which may be operating with a maintenance power level for their batteries. If Pgsc cannot be supported by the first chosen cluster, then the test is made for the sum of the first two clusters, with the worst SOC. If Pgsc cannot be supported by two clusters, then the test is made for the sum of the three clusters.

This testing loop aims to use only as many clusters as are needed (and no more - to minimise cluster power losses and battery cycling, and allow

background maintenance charging to continue on other clusters), and to preference using the cluster with the worst relative SOC level so that SOC levels attempt to equalise between clusters, that is, discharging events preferentially select the cluster with highest SOC as the cluster to use first - and charging events first select the cluster with lowest SOC.

For the decision branch where only one cluster is needed to meet the compensation power requirement, and the clusters have about the same SOC, then the cluster with the highest Pep level is chosen.

When multiple clusters are needed to provide the compensation power, then Pgsc is spread over the clusters with the same percentage loading as to Pcpi. When only one or two clusters are chosen to support Pgsc, then the choice of cluster can alternate when SOC changes. As an example, the SOC between the two clusters will be almost the same, and the choice of clusters may churn. A level of hysteresis (as an example, 2% of SOC) is used to bias the worst cluster so that particular cluster remains the chosen cluster to thereby avoid excessive churning.

For conditions where Pgsc >∑ Pcpi then all clusters are set at max Pcpi with PF=1 , as the fluctuation level is severe.

The SPS Power Limit Setting Required sub-process 51 is aimed to assist the GSCS 11 in controlling a rapid positive fluctuation in Ppv, by commanding a maximum power limit Ppvm (as a % of the maximum power limit level of Pgsc) to each inverter of the SPS 17, such that the maximum prospective power available from all the inverters of the SPS 17 is∑Ppvm = PRC.

Modifying Ppvm to minimise the occurrence and extent of non-zero values Pgsc, or in other words compensation power response support activity, is a generally preferred strategy.

Any future rapid increase in Ppv (as an example, due to a break in heavy cloud cover) can be constrained in magnitude to PRC-Ppv.

PRC is set to a suitable margin above Psra such that there is always 'headroom' above Psra for small fluctuations in Ppv, and rapid Ppv excursions above the deadzone limit are constrained.

This control technique:

• minimises the need for the battery to cover high-rate charging events.

• allows the quiescent battery SOC to be maintained at a higher level, as peak charging events are mitigated.

• allows the battery inverters to run cooler, as they don't experience peak charging events.

Table 3 directly below outlines this control strategy.

TABLE 3

If PRC < Psra +∑Pchi + Margin,

then transfer to the Set SPS Power Level sub-process 53. • ∑Pchi is the sum of cluster maintenance power levels Pchi and adjusts the test condition to cover maintenance charging when it is active.

• The Margin value is the headroom above the Psra level.

The Set SPS Power Level sub-process 53 uses the parameters PRC = ∑Ppvm, the individual SPS inverter power limit setting.

Table 4 directly below indicates how PRC is varied.

TABLE 4

If PRC < Psra +∑Pchi + Margin, then PRC is increased

If PRC > Psra +∑Pchi + Margin, then PRC is reduced

The capability of each cluster to generate a prospective power level is continuously calculated from parameters that include battery string SOC, the number of connected battery strings, battery string voltage, battery monobloc alarm levels, inverter temperature, inverter on-line status, and the polarity of the power.

Cluster power apportioning process

This is achieved through the Cluster power apportioning process 33, which involves SOC battery management and prospective cluster power processing. The cluster SOC will vary with each period where Pgsc≠0 or Pchi≠0 (ie. when Pci≠0).

The SOC needs to be maintained in a range that allows the GSCS 11 to adequately respond to future worst-case events, whether positive or negative going.

Empirical observation indicates that peak Pgsc activity tends to be higher when drops in Ppv occur (and the battery system 21 must discharge), rather than when increases in Ppv occur (and the battery system 21 must charge).

The capability of the battery system 21 to absorb charge(Pgsc<0) is less than its ability to discharge (PgsoO), as the charge efficiency falls rapidly above about 80-85% SOC, especially at high charge rates.

The likelihood of reduced battery service life increases as the time spent at lower SOC increases. The appropriate range of maintained SOC is a

compromise, and is presently considered to be 70-75%. The on-going calculation of SOC is prone to increasing error as time extends from when the last full equalise charge was undertaken (at which time SOC level was reset to the known level of 100%). On-going calculation errors are incurred from kWh measurement; self discharge rate as a function of SOC and temperature; charge and discharge inefficiency as a function of power, SOC and temperature; accuracy of the compensation equations used by the PLC.

To maintain the SOC, the PLC system 23 uses two sub-processes: a Calculate string SOC sub-process 55 and a Set Maintenance Charge Level sub- process 57.

Calculate string SOC sub-process

The Calculate string SOC sub-process 55 calculates SOC by tracking battery capacity using normalised capacity changes each second. The cluster power is divided amongst the on-line battery strings, and any power losses in transferring cluster power to the battery strings are also apportioned. The cluster power level Pci is referenced to the AC port of the inverters, and so incurs additive/subtractive losses when referenced to a battery depending on charge or discharge power flow. The battery string SOC, SOCj, is reset to a limit level during a maintenance charge (for example, 100% for full equalisation). Table 5 directly below outlines this.

TABLE 5

SOCj = 100^*Cbatj / Cbatjr (%)

Cbatj = Cbatjt-1 - KTj^*KRj^*KSj^*(Pci^*1000^*%j + PLi)/3600

Cbatjr is the maximum string capacity in Wh at T=20°C, and at a 5 hour discharge rate, for battery string j. Cbatjr is modified over time as battery strings age, based on periodic full discharge capacity tests of each string, but is taken as a constant in between periodic tests.

Cbatj is the actual string capacity in Wh of battery string j, when normalised to T=20°C, and a 5 hour discharge rate. Cbatj increases when the charging power level (negative value) exceeds the parasitic loss PLi. When Pci=0 then there is a parasitic loss PLi in string capacity. PLi is the sum of the inefficiency contributions from inverter loss, internal battery and cable resistances, and charging, and is normalised back to an on-line per battery level. Resistance and charging losses are a function of Pci, and charging loss is also a factor of SOC (eg. charging loss is 1 00% when the battery reaches SOC=100%). Constant losses are inverter loss, and monitoring loss. Losses at the cluster level are divided by the number of battery strings connected, BatOK, and only applied to battery strings connected.

Table 6 directly below outlines how inefficiency contributions are calculated.

TABLE 6

PLi = (PLLi/BatOK) + PLRi

Total cluster inverter losses are PLLi = 3 x 80W + Pci^*1000/(25.4), when activated.

Inverter efficiency is 96%; no-load active loss is 80W. PLLi = 6W when inverter is not activated.

Battery string resistance = Rb, and battery voltage = Vb.

Battery string resistive loss PLRi = (Rb/BatOKi) ^* (Pci^*1000/Vb)² per battery per cluster.

KTj translates battery string operation at a temperature TBj different from the reference level of T=20°C, which accounts for dependence upon actual temperature. Table 7 directly below indicates how KTj is calculated.

TABLE 7

KTj = 0.0002^*(TBj)₂ - 0.01 32^*TBj + 1 .1 98 ; KTj=1 for T=20°C

KRj translates battery string operation at a discharge rate (Rate), which different from the reference level of a 5 hour discharge rate. Table 8 directly below indicates how KRj is calculated.

TABLE 8 For PckO: KRj = 0.08^*Rate +0.92 ; Rate=Pci^*1000/(BatOK^*Cbatjr/5) For Pci>0: KRj = 1

Charge acceptance is effectively 1 for SOC up to 75-80% and then drops, and this is accounted for in KSj.

When Pgsc = 0 and Pchi=0 with the cluster activated, then Pci is set at a residual low level to offset PLLi, to operate the batteries at open-circuit condition. Set maintenance charge level sub-process

The Set Maintenance Charge Level sub-process 57 calculates a maintenance power level Pchi for each cluster i. Pchi is used in the Set Cluster Power Level sub-process 49.

Maintenance Power level Pchi:

• Pchi = Fchi ^* Fchsi ^* Pchr

Maintenance power of a cluster continues as long as the cluster is not required to support a Pgsc≠0 event.

Maintenance power required factor Fchi:

· This factor switches on the maintenance power, when SOC is outside a range of SOC centred on StartSOC, with the range set by SOCrange.

• If SOCckStartSOC -SOCrange then Fchi=-1 ; OR

• If SOCc StartSOC +SOCrange then Fchi=+1 ; OR

• Otherwise Fchi=0.

· If EQi=1 , then this sub-process is bypassed and Fchi=-1 is forced to a maintenance power to equalise the battery.

• If EQDi=1 , then this sub-process is bypassed and Fchi=1 is forced to maintenance power to discharge the battery.

The smoothing ramp factor Fchsi provides a ramp from 0 to 1 , or 0 to -1 , over a duration to smooth the application of the maintenance power.

The Maintenance Power Level Pchr is selected to a setting suitable for the battery.

Prospective cluster power processing provides for the maximum prospective throughput power of a cluster, Pcpi, which is set by the inverter activate status of the bidirectional inverter system 19, inverter thermal limits, battery SOC, and the number of battery strings connected. The inverters of the bidirectional inverter system 19 have power limited throughput based on internal inverter temperatures.

Battery SOC is continuously calculated by the PLC system 23 for each battery string and cluster of the battery system 21. A factor Fbtj modifies Pcpi to account for SOC with both charge and discharge conditions. A battery string status factor Fbvj identifies when a string is not connected.

Cluster capability sub-process

Prospective cluster power processing uses a Cluster Capability sub- process 59 that determines the parameters Pcpi, which constitute the individual cluster prospective power levels.

The cluster power capability is the minimum capability of the 3 inverters in the cluster, as cluster output power must be balanced.

Batteries are considered able to meet any demand, however a derating is introduced towards the limits of the acceptable operating SOC range. Derating starts at <24% SOC and >83% SOC. Derating is selectively applied to negative and positive demands - in other words, a low SOC causes derating of positive Pep, whereas a high SOC causes derating of negative Pep.

It should be appreciated that the characterising equations used are only exemplary and may change based on test performance of a particular site.

Thus, a compensation power level response is apportioned to one or more clusters according to the polarity and magnitude of the response level; the prospective power level capability of each cluster; whether a cluster is previously supplying a level of compensation response; and whether a cluster is previously maintaining its battery strings.

The DPS 16 may command a maximum level of power generation from the site power source 15 for reasons which include appropriate diesel genset loading conditions and allocation of defined levels of spinning reserve.

The PLC system 23 transfers the maximum power level to the site power source 15 power level to the SPS 17. In addition, the PLC system 23 may further reduce the maximum SPS 17 power level to within a controlled margin above the present site power source 15 output power level. The combination of a controlled margin, and a ramp rate increase characteristic for the power limit level of the SPS 17, can minimise the requirement for the GSCS 11 to compensate for increasing SPS 17 power levels.

In addition, the PLC system 23 may further control and limit maximum power level of the SPS 17 for the purpose of full or partial shutdown of the site power source 15. In addition, the PLC system 23 may further control and limit maximum power level of the SPS 17 in response to a full or partial reduction in prospective power level capability of the GSCS 11.

SPS power command process

The power capability of Ppv can be controlled by the PLC system 23 using the SPS power limit command process 35, which transfers the command setting to each inverter of the SPS 17. This process is required for the following sub- processes to be performed:

• SPS Power Limit Setting Required sub-process 51 to set the maximum power setpoint.

· PLC command to manage when Pep does not adequately cover PRC.

• Set SPS Power Limit Level sub-process 53, which manages rapid large positive fluctuations in Ppv.

• PLC command to manage shutdowns, which is performed as part of the Operating Mode process 57.

The DPS 16 can issue a SPS Max Power Setpoint at any time.

Whenever SPS Max Power Setpoint > PRC, then this process has no further outcome.

Where SPS Max Power Setpoint is below PRC, then the PRC is reduced until PRC < SPS Max Power Setpoint.

During a DPS-commanded reduction in SPS Max Power Setpoint, Ppv will ramp down at fastest rate configured for PV inverters due to lower PRC, and communications time delays to SPS inverters.

The GSCS 11 can vary PRC if needed, as long as PRC < SPS Max Power Setpoint.

Operating mode process

The activation and deactivation of clusters during the day is made a function of the SPS 17 output power level, the SPS 17 power limit level PRC, and the prospective power level capability of activated clusters Pcpi. The number of activated clusters is kept to a minimum to reduce power loss within the site power source 15.

When a cluster is activated, the cluster's battery strings are maintained at open-circuit conditions unless there is a need for the cluster to generate a compensation power response, or a maintenance power response. The cluster's battery strings are maintained within a target range of partial state-of-charge by calculating the change in state of charge and tracking the state of charge and applying a maintenance charge or discharge.

The Operating Mode process 37 controls the activation and deactivation of clusters using a Cluster Activation sub-process 61. As a normal function, clusters are activated using the MODEi parameter, which is based primarily on the level of Ppv. Cluster i can be turned off at any time for the purpose of protection by setting the parameter CLiEN = 0.

Activation preferences cluster 1 , then 2, then 3, but manages when Pcpi of a cluster is low. Table 9 directly below indicates the conditions under which MODEi is active.

TABLE 9

MODEi is active when: : Ppv > LIMneg + Z2^*∑Pcpi

De-activation preferences cluster 1 , then 2, then 3, but manages when Pcpi of a cluster i is low. Turn-off is delayed for 15 minutes after the turn-off conditions are met, to avoid churn.

The application of a maintenance power level to a cluster in the Set Maintenance Charge Level sub process 57, and the setting of Pci to a residual low level when Pgsc=0 and Pchi=0, combine to apply intermittent charging of the cluster batteries, with the batteries operating under open-circuit conditions in a partial state of discharge when the maintenance charge process is not active.

The Operating Mode process 37 includes discrete sub-processes to perform other functions, namely: Site Power Source Shutdown sub-process 63 and DPS Directed Start & Start sub-process 65. In the Site Power Source Shutdown sub-process 63, DPS-directed shutdown of the site power source Hard Shutdown is activated by the DPS STOP command.

The PLC system 23 executes the following procedures:

· Set SPS Urgent Alarm, and set SPS Stop Enabled status=1 .

• Set SPS inverter Ppvm to 0%.

• Set Pgsc=0.

• Set CLiEN=0 for all clusters i.

The site power source Limited Shutdown is activated by DPS command error, DPS communications alarm, and Meter communications alarm. The PLC system 23 executes the following procedures:

• Set SPS Urgent Alarm and set SPS Start Enabled status=0.

• Lower PRC at LI Mneg/minute rate.

• Set CLiEN=0 and SPS Stop Enabled status = 1 when Ppv =0.

The site power source Emergency Shutdown is activated by a single

Emergency Stop button in the Main Switchboard of the GSCS 11 by opening all the main switch circuit breakers in the Main Switchboard.

During a shutdown the AC and DC power to the cluster inverters of the bidirectional inverter system 19 are isolated.

With the DPS Directed Start & Stop sub-process 65, the DPS issues a

START and STOP command to the site power source 15.

After a START condition is identified, and acknowledged with SPS Start Enabled status = 1 level, the GSCS 11 can operate normally.

After a STOP condition is identified, the PLC commands a Hard shutdown as described previously.

Remote control process

The Remote Control process 39 allows communication with the PLC system 23 from offsite. The communication set up of such is common knowledge in the art and will not be described further. As previously mentioned, the remote control process 39 provides for:

• responding to remote start and stop commands, and control levels;

• responding to remote communications in relationship with watchdog and failure conditions; and • providing operational process levels to the remote communications device.

Battery maintenance process

The Battery maintenance process 41 involves maintaining each battery string by:

• A Battery Contactor sub-process 67.

• A Battery Equalisation Charge sub-process 69.

• A Battery Discharge sub-process 71.

Each battery string is monitored by a pcb with:

· Wired voltage free contacts connecting to PLC Dl's for High Voltage

(HV), Low Voltage (LV), a common alarm, and contactor status. The common alarm combines Overvoltage (OV), Undervoltage (UV), Overcurrent (OC), and Earth Leakage (EL) alarms.

• Wired PLC relay contact DO's for Contactor Enable and Contactor Latch.

• Other connections for custom manual monitoring, and to support the pcb (power and sense cabling).

The HV and LV Dl's are used to initiate process transitions that terminate or derate a particular battery string during a charging or discharging event that is being controlled by the PLC.

The pcb de-energises the contactor for OV and UV conditions after alerting the PLC via the common alarm, which disconnects the battery string from the inverter.

The PLC system 23 can force a contactor on, so the alarm can be overridden if needed, by forcing an Enable and a Latch on the contactor.

Battery contactor sub-process

In the Battery Contactor sub-process 67 involves each battery string including a contactor that is under the control of the PLC and the local battery monitor pcb as previously described.

During normal daily operation all battery string contactors are closed, but would be sequentially commanded open when LV alarm becomes true, except that when the last battery string presents an LV alarm then the cluster Pcpi is set at zero to inhibit PckO. During normal daily operation all battery string contactors are closed, but would be sequentially commanded open when HV alarm becomes true, except that when the last battery string presents a HV alarm then the cluster Pcpi is set to zero to inhibit Pci>0.

The contactor auxiliary status feedback is used to modify BatOK parameter for SOC level caclulations and for cluster prospective power level calculations. Battery equalisation charge sub-process

The Battery Equalisation Charge sub-process 69 forces the battery strings in a cluster through an equalisation charge, which is required to periodically reset the SOC tracking level, and also to maintain the battery health. The process requires a remote command signal RSi_Start to initiate, and time-of-day constrained to only start between a prescribed time of the day, and only when Pci=0 and all battery strings are connected. The process ends the following day at a prescribed time, after which the SOC is reset to 100%, and the maintenance power process may then change the SOC to the target range.

The remote command signal RSi_Start is active throughout the process, and can be used to deactivate the process. The RSi_Start_Enabled flag remains active during the process, to identify that the process is active, and is reset to zero at the prescribed time when the process is terminated.

The Calculate Individual Inverter Demand sort process is biased to not use a cluster undergoing an equalisation charge process, however if the other available cluster(s) cannot fully support a compensation power level, then the equalisation charge process is terminated and that cluster is used to assist supporting the compensation power level.

If a common alarm is raised after an LV or HV has been raised, then the cluster is disabled by setting CLiEN=0.

Battery discharge sub-process

The Battery Discharge sub-process 71 discharges the battery strings in a cluster and allows the pre-existing capacity of the cluster batteries to be measured. The pre-existing capacity can be used to calculate the actual preexisting SOC level, for comparison with the running prediction of SOC level at that time. The process requires a remote command signal RSDi Start to initiate, and time-of-day constrained to only start between a prescribed time period, and only when Pci=0 and all battery strings are connected.

The remote command signal RSDi Start is active throughout the process, and can be used to deactivate the process. The RSDi_Start_Enabled flag remains active during the process, to identify that the process is active, and is reset to zero when the process is terminated.

Sequential charge and discharge processes 69, 71 allow the full-discharge capacity of the cluster batteries of the battery system 21 to be measured.

Comparison of benchmark full-discharge capacity levels over many years can assist in managing the programmed replacement of the batteries.

The GSCS 11 operation is biased to not use the battery cluster undergoing a Battery Discharge sub-process 71 , however if the available battery clusters cannot fully support a compensation power level, then the Battery Discharge sub- process 71 is terminated and that battery cluster is used to assist supporting the compensation power level.

Facility control process

The Facility Control process 43 involves various PLC relay contacts for local annunciation of events, local activation of a process, local power cycling of communications equipment.

The Facility Control process 43 involves several sub-processes, including a Facility Monitoring sub-process 73, a Data Logging Parameters sub-process 75, a Temperature Measurement sub-process 77, and a Reactive Power Control sub- process 81.

The Facility Monitoring sub-process 73 monitors the doors to the compartment containing the GSCS 11. A non-urgent alarm is raised when a door is open. The doors need to all be closed for normal operation.

The Control 24VDC supplies are monitored for OK status. A non-urgent alarm is raised when any one of the three redundant supplies is not OK.

Data logging parameters sub-process

The Data Logging Parameters sub-process 75 calculates and updates specified parameters to the DPS 16 at required intervals, and made available for monitoring and reading.

Temperature measurement sub-process The Temperature Measurement sub-process 77 reads various temperature sensors in the battery compartment. Each battery inverter has a temperature sensor attached to each battery string. Measured battery compartment temperature is read from each bidirectional inverter 19.

Air-conditioner control sub-process

The Air-conditioner Control sub-process 79 operates air-conditioners in the compartment to constrain battery inverter internal temperature to within equipment maximum operating temperature limits, and also to maximise Cluster capability during high PRC periods by retaining inverter temperature at 25°C.

The air-conditioners use permissive contact activation from the PLC via a solid-state relay powered from 24VDC.

Room temperature control is achieved by control via the PLC system 23 of the number of active air-conditioners. The number of operating air-conditioners is increased with cumulative inverter thermal stress. Inverter thermal stress is related to inverter fan speed Fsk and inverter thermal temperature Tlk.

The control demand parameter AIR determines the number of air- conditioners enabled, as indicated in Table 10 directly below.

TABLE 10

AIR =∑(FSk) +∑(FSk>0.35) + IF(MIN(Tlk)>30+ToD,0.5,0) ,

where ToD=0 for >9am, <4pm; and ToD=15 for <9am and >4pm. AIR = 0 then all aircons disabled;

0 < AIR <1 , then one aircon enabled

1 < AIR <2, then 2 aircons enabled;

... AIR >3, all aircons enabled. Disabling occurs when AIR falls below enable range, plus 2 minute delay.

The contribution of ∑(FSk) represents the cumulative dissipation of all inverters. The contribution of ∑(FSk>0.35) represents the number of stressed clusters. The contribution of Tlk represents the compartment air temperature.

Reactive power control sub-process

The Reactive Power Control sub-process 81 allows the PLC to control the power factor PF of each cluster. The default PF level is 1 .0. If PF is not set to 1 .0, the PF level reverts to 1 .0 when the demand level of Pgsc exceeds the cluster capability level∑Pcpi.

The operation of the communication interfaces, input handling and the programming required to achieve same is standard engineering knowledge to the skilled person in the art and will not be described further.

It should be appreciated that the scope of the present invention is not limited to the preferred embodiment described in relation to the best mode for carrying out the invention.

The invention is not limited to use of photovoltaic-based power generation, and other forms of supplementary power sources may be used, such as wind or tidal power.

Alternative battery configurations and technologies such as single string configuration and Ni-Cad, Li-Ion technologies may be used.

A range of alternative controller technologies may also be used, such as microprocessor control embedded in to a personal computer or an inverter, or multiple control devices may be used.

The invention is not limited to three single-phase inverters in three-phase form of bi-directional inverter, and other configurations such as single-phase form and three-phase inverter in three-phase form may alternatively be used.

While a one-minute rolling average form of reference variable is described, and other forms of reference variable, such as involving different time spans, and sample rates or weighting variables may be used to suit local conditions and selected application.

Similarly, while a constant ramp limit form of rate of change limiting is described, a variety of other rate of change limit characteristics may be selected, such as stepped rates or conditional rates, or other more complex contraints.

Other deadzone limit characteristics such as stepped or conditional limits may also be used as required.

Claims

1. A method of controlling the stability of an electricity grid that is supplied electrical power from a relatively unstable power source, which is supplemented by a battery system and a bi-directional inverter system to provide a combined power output, the method comprising steps performed at a controller of: receiving samples of a real-time power output of a relatively unstable power source; calculating a real-time parameter as a smoothed representation

characteristic of recent values of the sampled real-time power output of the relatively unstable power source; determining whether the sampled real-time power output of the relatively unstable power source exceeds predetermined negative or positive limits from the smoothed representation of the sampled real-time power output of the relatively unstable power source; and commanding a compensation power response by discharging or charging the battery system via the bi-directional inverter system to mitigate a disruptive change in the power output of the relatively unstable power source, when the measured power output is determined to diverge from the smoothed

representation by either a respective predetermined negative power limit or a predetermined positive power limit.

2. A method according to claim 1 , wherein a rolling average of the combined power output is used as the smoothed representation characteristic of the recent power output of a relatively unstable power source.

3. A method according to claim 1 , wherein a rolling average of immediately preceding values in the range of 30s to 90s of the combined power output is used as the smoothed representation characteristic of the recent power output of a relatively unstable power source.

4. A method according to claim 1 , wherein the predetermined negative and positive power limits are calculated as a predetermined percentage of

predetermined absolute limits.

5. A method according to claim 1 , further comprising continuing to provide a compensation power response until the sampled real-time power output of a relatively unstable power source has returned to within a predetermined margin of a rolling average of the combined power output.

6. A method according to claim 5, wherein the predetermined margin is calculated as a predetermined percentage of the smoothed representation.

7. A method according to claim 1 , wherein the combined power output of the relatively unstable power source and the compensation power response is controlled to a maximum predetermined negative rate of change, and a maximum predetermined positive rate of change.

8. A method according to any one of claims 1 to 7, wherein the primary power source is a diesel generating power source, and the relatively unstable power source is a power generating source selected from the group comprising: a solar power generating source, and a wind power generating source.

9. A method of controlling the stability of an electricity grid that is supplied electrical power from a relatively unstable power source, which is supplemented by a battery system and a bi-directional inverter system to provide a combined power output, the method comprising steps performed at a controller of: receiving samples of a real-time power output of a relatively unstable power source; calculating a real-time parameter as a smoothed representation

characteristic of recent values of the sampled real-time power output of the relatively unstable power source; determining whether the sampled real-time power output of the relatively unstable power source deviates from recent values of the sampled real-time power output of the relatively unstable power source; and commanding a compensation power response by discharging or charging the battery system via the bi-directional inverter system to mitigate a disruptive change in the power output of the relatively unstable power source; wherein the combined power output of the relatively unstable power source and the compensation power response is controlled to a maximum predetermined negative rate of change, and a maximum predetermined positive rate of change.

10. A method as claimed in claim 9, wherein the commanded compensation power response is calculated as a function of the difference between the power output of the relatively unstable power supply and a rolling average of the combined power output.

1 1 . A method as claimed in claim 9, wherein the commanded compensation power response is calculated as a function of an incremental difference in a rolling average of the combined power output, multiplied by the maximum predetermined negative rate of change or maximum predetermined positive rate of change.

12. A method as claimed in claim 9, further comprising determining whether the sampled real-time power output of the relatively unstable power source deviates from the smoothed representation of the sampled real-time power output of the relatively unstable power source by one of a predetermined positive power limit and a predetermined negative power limit, wherein the compensation power response is commanded, when the measured power output is determined to exceed one of the predetermined positive or negative power limits.

13. A method as claimed in claim 9, wherein the maximum predetermined positive rate of change and maximum predetermined negative rate of change are calculated based upon, respectively, the predetermined positive power limit and predetermined negative power limit.

14. A method of controlling the stability of an electricity grid that is supplied electrical power from a relatively unstable power source, which is supplemented by a battery system and a bi-directional inverter system to provide a combined power output, the method comprising steps performed at a controller of: receiving data representative of state of charge of each of a number of clusters of a multi-cluster battery system; calculating a maximum prospective power of each of the battery clusters; and commanding selective charging and discharging of the battery clusters to conform to maintain the battery clusters within a target range of measured state of charge.

15. A method according to claim 14, further comprising preferentially discharging the battery clusters with highest state of charge is performed with a predetermined hysteresis margin to reduce churning between battery clusters.

16. A method according to claim 14, further comprising commanding the battery clusters to remain in open circuit mode when not discharging.

17. A method according to claim 14, wherein the target range of state of charge is in the range 70 to 75% state of charge.

18. A method according to claim 14, further comprising commanding periodical equalisation charge of each of the battery clusters.

19. A method according to claim 14, further comprising commanding selective recharging each of the battery clusters to a target state of charge in the range 70 to 75% state of charge.

20. A controller for controlling the stability of an electricity grid that is supplied power from a primary power source, and a relatively unstable power source, and which is supplemented by a battery system and a bi-directional inverter system, the system comprising a controller for performing steps according to any one of claims 1 to 13.

21 . A power station for supplying electrical power comprising a primary power source, and a relatively unstable power source, which is supplemented by a battery system, a bi-directional inverter system, and a controller, wherein the power generated by the relatively unstable power source is monitored, and a compensation power response provided via the battery system and bi-directional inverter under action of the controller performing steps according to any one of claims 1 to 13.

22. A controller for managing state of charge of a battery system used in conjunction with a bi-directional inverter system used for compensating for fluctuations in power output by a relatively unstable power source, the controller comprising means for performing steps according to any one of claims 14 to 19.

23. A power station for supplying electrical power, which is supplemented by a battery system, a bi-directional inverter system, and a controller, wherein the state of charge of the battery system is managed by a controller performing steps according to any one of claims 14 to 19.

24. A system for controlling the stability of an electricity grid that is being supplied electrical power from an RU power source, where the RU power source is supplemented by a battery and inverter system, comprising a controller operating a plurality of processes including a rate of change output power limiting process to accommodate the electrical grid, the rate of change output limiting process including: a power response control process for calculating compensation power command levels; and a power command process for controlling the power command level to the RU power source; wherein a compensation power response is provided to partly mitigate against disruptive change provided by the power command level by adjusting the rolling average real power output to the electrical grid to ramp at a maximum allowable ramp rate.

25. A system according to claim 25, wherein the power response control process involves:

• monitoring of the RU power source power generation level; · testing for compensation power requirement;

• calculating compensation power level; and

• apportioning compensation power level to cluster power levels.

26. A system according to claim 25 or 26, wherein the power command process involves: · responding to external control power limit requirements;

• calculating an RU power source power limit level margin;

• communicating the RU power source power limit level; and

• monitoring the RU power source power limit level.

27. A system according to any one of claims 25 to 27, wherein the rate of change output power limiting process includes a cluster power apportioning process for apportioning the compensation power levels to cluster power command levels.

28. A system according to claim 27, wherein the cluster power apportioning process involves:

• monitoring cluster power limitation parameters;

• calculating prospective cluster power capability; · biasing apportioning calculations to reduce cluster churn and

maintain maintenance processes; and

• apportioning compensation power level.

29. A system according to any one of claims 24 to 28, wherein the rate of change output power limiting process includes an operating mode process for energising and de-energising the battery and inverter system.

30. A system according to claim 29, wherein the operating mode process involves:

• activating and deactivating clusters to support peak prospective compensation power requirement; · activating and deactivating battery strings of the battery and inverter system to support operational capability of each battery string; and

• operating the battery strings in a partial state of discharge and in open-circuit conditions.

31 . A system for controlling the stability of an electrical grid that is being supplied power from an RU power source, where the RU power source is supplemented by a battery and inverter system, comprising a controller operating a plurality of processes including a deadzone control process to maximise battery open circuit mode duration and minimise the use of inverters, the deadzone control process including: • a test for active response level sub-process to read the total real power parameter of the RU power source;

• an active response required sub-process to signify deadzone power limit levels, the power limit levels being set to alleviate the need for the system to act on small changes in power output of the RU power source, and determine whether a compensation power response is required having regard to the total real power parameter of the RU power source and the deadzone power limit levels;

• a set cluster power level sub-process to provide the power compensation power response using the battery and inverter system in a manner to minimise power losses and battery cycling.

32. A system according to claim 31 , wherein the deadzone control process includes a power setting required sub-process to assist in controlling a rapid positive fluctuation in the power level of the RU power source by commanding a maximum power limit as a percentage of the maximum power that can be transferred through the battery and inverter system.

33. A system according to claim 31 or 32, wherein the set cluster power level sub-process also operates to maintain background maintenance charging of battery clusters where required.

34. A system for controlling the stability of an electrical grid that is being supplied power from an RU power source, where the RU power source is supplemented by a battery and inverter system, comprising a controller operating a plurality of processes including: (i) an immediate compensation response to compensate for large dynamic transient disturbances in the RU power source; and (ii) a slow compensation response to compensate for slower fluctuations.

35. A system according to claims 34, wherein the immediate compensation response is triggered to mitigate a disruptive change in the power output of the relatively unstable power source, when the measured power output is determined to diverge from recent values of power delivered by the RU power source by either a respective predetermined negative power limit or a predetermined positive power limit.

36. A system according to claim 35, wherein a rolling average of immediately preceding values in the range of 30s to 90s of a combined power output of the power delivered by the RU power source and the compensation power response is used as the smoothed representation characteristic of the recent values of power delivered by the RU power source.

37. A system according to claim 36, wherein a compensation power response is provided by an immediate power response or a slow power response until a sampled real-time power output of the RU power source has returned to within a predetermined margin of the combined power response.

38. A system according to claim 34, wherein a combined power output of the RU power source and the immediate or slow power response is controlled to a maximum predetermined negative rate of change, and a maximum predetermined positive rate of change.

39. A system for controlling the stability of a grid that is being supplied power from a RU power source, where the RU power source is supplemented by a battery and inverter system, comprising a battery manager operating a plurality of processes including a partial state of discharge targeting process to manage the SOC of a plurality of battery clusters comprising the battery and inverter system.

40. A system according to claim 39, wherein at least one of the processes of the battery manager receives data representative of state of charge of each of the battery clusters of the battery and inverter system.

41 . A system according to claim 39 or 41 , wherein at least one of the processes of the battery manager calculates a maximum prospective power of each of the battery clusters, and commands selective charging and discharging of the battery clusters to conform to maintain the battery clusters within a target range of measured state of charge.

42. A system according to any one of claims 39 to 41 , wherein the battery clusters having the highest state of charge are preferentially discharged with a predetermined hysteresis margin to reduce churning between battery clusters.

43. A system according to any one of claims 39 to 42, wherein the battery clusters are controlled to remain in open circuit mode when not discharging.

44. A system according to any one of claims 39 to 43, wherein the target range of state of charge of each of the battery clusters is in the range 70 to 75% state of charge.

45. A system according to any one of claims 39 to 44, wherein at least one of the processes commands periodic equalisation charge of each of the battery clusters.