What will microgrids mean for Aussie electricians?
Microgrids have grown considerably since they were ushered in nearly a century ago. Phil Kreveld looks at where they stand and what it’ll do for the Australian electrical market.
Once upon a time, when electric power was new to the world, there were only microgrids; Edison’s New York and Michael de Ferranti’s London. When alternating current won the toss from direct current, it became practical to extend networks because of transformers and therefore high voltage to pipe large amounts of energy to many, many consumers.
In essence, nothing much has changed since the 1900s—the same type of ‘dynamos’ (synchronous generators), transformers, switchgear, etc. serve us today. The legacy technology from yesteryear is now under pressure because of consumers taking care of their own energy needs by way of rooftop solar.
One-way power flow is now two-way, pushing power into the transmission grids as a result of people buying larger and larger rating solar panel-inverters. Australia is the world’s standout because of the copious availability of sunlight and consumers, who not only want to reduce their electricity bills but want to make money from exported power. The Australian Energy Market Operator (AEMO), in its step-change scenario for the adoption of renewables, estimates domestic solar to have an aggregate capacity of 18GW within five years, amounting to 75% of demand of the South East coast network.
Something has to give
A great deal of effort is going into all kinds of marketing schemes to make domestic solar system owners happy, for instance by not curtailing their power generation, by rewarding them via aggregators for frequency support, battery rebates and community batteries.
Something has to give in the end because all those domestic solar inverters require a stable source of voltage and frequency and that requires a minimum level of power inflow into distribution networks. Western Australia has already experienced this situation, urging consumers to keep their air conditioners switched on to provide sufficient inflowing load. All domestic solar inverters are grid-following and their anti-islanding protection plus limited low voltage ride-through make them unsuitable to operate independently.
Off the grid
Yackandandah and Daintree have become emblematic of microgrids, but in fact, they are nothing new in that there are plenty of isolated grids in Western Australia where they are the only sensible alternative to extending high voltage lines over hundreds of kilometres. Diesel generator systems are an efficient solution for small grids but the pressure on CO2 reduction is popularising the use of battery energy supply systems (BESS).
For the moment we have no examples of DC microgrids, although they can be a practical solution. For AC reticulation, voltage forming inverters have to be connected to batteries. These are in smaller ratings, typically under 50MW, commercially available technology, for example, the 30MW, Hitachi-ABB Eyre peninsula microgrid. A pesky problem, although resolvable over time, is getting more than one voltage forming inverter to power-share, whereas that is a natural feature of synchronous generators. Sudden spurts in power can also pose a problem for battery supported inverters because of their limited over-current ratings, typically between 120% and 150% compared to 500% for synchronous generators.
Micro (macro) grids within distribution grids
Separable portions in distribution grids seem a farfetched idea. Yet such a form of microgrid or ‘macrogrid’ has advantages in the sort of networks we are heading for. AEMO is having to intervene in the energy market at the rate of several hundred instances per annum in order to preserve stability.
Demand control, by putting a price on demand reduction, could be replaced by carving off demand in distribution networks thus creating microgrids capable of sustaining demand within themselves. In fact, taking account of the growing household solar capacity to which can be added commercial and industrial capacity, the National Electricity Market (NEM) is heading towards the self-sufficiency of many zone substations (macrogrids).
Based on BESS-inverters, such independence of HV transmission power would require voltage forming inverters with appropriate droop control to permit power-sharing. Currently, huge battery systems are being constructed to provide grid support in the HV transmission system, yet these could also be employed at the distribution level provided their inverters had voltage forming properties.
Grid supporting inverters and grid forming inverters seem like the same thing but that is not necessarily the case. The Basslink DC cable carrying energy from Tasmania to the mainland terminates at Loy Yang power station.
There, the DC power is converted to AC by inverters that depend on AC voltage for the commutation of the power semiconductor components. The inverters ‘support’ but cannot initiate the grid, for example in a black start situation. Grid forming, on the other hand, implies voltage forming, and therefore in principle, black start capability. In figure 1a, a grid following inverter is shown, and in figure 1b, a voltage forming inverter.
Grid following inverter. Power, P, and reactive power, Q, values are inputs into the control loop, altering the current i magnitude and its angle with respect to the grid voltage. Although the inverter is connected to a DC voltage (i.e., is voltage sourced) it is the current which is controlled. Such an inverter is often referred to as a current source.
Voltage forming inverter. The input variables are frequency, ω = 2πf, and voltage, E. The inverter is seen as pure voltage source with a series impedance connecting it to the grid. The voltage, v, amplitude and its angle viz a viz the grid voltage vary. The voltage angle is important as it relates to the real power, P, delivered as determined by the equation where E is the voltage amplitude required, V is the grid voltage and d is the angle between E and V, and X is the impedance (assumed as inductive) of Zeq. Frequency in effect, determines the angle, d, the ‘power angle’ according to a droop characteristic,
In the droop equation, f, is the measures frequency, f0, is the control frequency, P and P0, respectively actual power and required power. The angle, d, is varied in accordance with the droop function. E, as is the case for synchronous generators, relates to the reactive power required by the load, and has its own droop characteristic as shown:
with Q and Q0 being the respective reactive power demanded by the load and the reference value.
Forms of microgrids
The simplest form of microgrid comprises of one power source, in the case of a BESS, a three-phase voltage forming inverter providing a stable voltage and frequency source for many, likely single-phase, grid following inverters of individual consumers and passive loads. The grid (voltage) forming inverter, as explained in figure 1b, behaves as would a synchronous generator, adjusting power output so as to keep frequency constant and reactive power in order to provide constant voltage.
Note that the assumption is that the feeder link to the load is mainly inductive. The BESS requires its own energy source in order to recharge it regularly. In the case of several BESS, operating in parallel, the droop characteristics as per figure 1b, will provide for power and reactive power sharing between them.
Series grids are more complicated in the power sharing aspect, and are not likely to be encountered other than as theoretical examples. In series grids, as the name implies, loads and generators are daisy-chained and there is no primus inter pares (a first amongst equals). This makes for either the need for a separate communication network between generators or a more complicated power sharing scheme.
Community batteries are being politically promoted and there are some practical trials being undertaken. They are seen as a means of avoiding or limiting reverse power flow and at the same time providing stored energy for households when tariffs are high while saving households from having to invest in as yet expensive batteries. They could be part of sustaining a microgrid but not without controllers and a form of inverter currently not available although the technology is simple enough in theory.
The inverters for community batteries are two-way power converters, and this function is basic to bridge-type power electronics, whether single- or three-phase. In the rectifying mode the triggering of the insulated gate bi-polar transistor (IGBT) switches is achieved by the line voltage. When operating as inverters, i.e., supplying energy to households, they use the line voltage to synchronise their pulse width modulation (PWM) to provide AC current. Synchronisation is achieved by the inverter phase locked loop (PLL) circuit, basically a voltage-controlled oscillator with a phase comparator input. Were the community battery inverters to also operate in a micro grid, the PLL would be disabled in favour of the voltage-forming mode. Modulation is then via the internal oscillator, previously synchronised by the PLL and as explained above, subject to droop control.
Microgrids and AEMO
AEMO does not raise the subject of microgrids within distribution grids. That said, AEMO is plenty-worried about distribution problems. For example, a MV fault interrupts power, solar inverters go into anti-islanding mode, switching off and when power is restored, there is a period of several seconds or longer, during which solar inverters are resynchronising to the grid. In this interval the substation load rises from previous low net power inflow when solar systems were operating full pelt, to a high value. The sudden power inrush required from synchronous generators, maybe hundreds of kilometres distant, momentarily lowers the frequency and as a result the distribution network under-frequency load-shedding relay (UFLS) isolates the network.
Another scenario also involving the operation of the UFLS relay, is during reverse power flow, increasingly prevalent. An under-frequency event occurs in the grid, and UFLS operation now exacerbates the situation by dropping off solar generation supplying power to the grid! These are not theoretical events—they actually occur!
Carving out grids within grids
The simplest example of this process is the uninterruptible power supply (UPS). The automatic transfer switch ensures that the diesel generator is in synch with the grid when power is restored, while during grid failure either batteries and inverters or a motor-generator as in a Diesel Rotary UPS (DRUPS) supplies power to essential loads.
Extending this to a distribution network might require the identification of essential loads. However, household loads would have to be regarded as essential and the automatic transfer switch is now replaced by control systems including black start. Of course, the latter may be avoided in a ‘smooth’ transfer from mains grid to microgrid when, for instance, an overloaded interconnector might require distribution grids to ‘go it alone’ but in all probability, a black start would be required after isolation of the distribution grid.
Black start more than likely has to be achieved by ramping, certainly for voltage forming inverters as their limited short circuit capacity of typically less than 150% of full load is no match for charging cables and establishing magnetic flux in transformer cores being fed from their secondary windings. BESS-inverters will be required to provide for power-sharing between each other via droop control. In contrast to synchronous generators, equipped with governors, the frequency determined by power-sharing is likely to be different from synchronous frequency, therefore requiring a master BESS-inverter to provide the 50Hz reference.
Power-sharing and synchronisation in the absence of a single source capable of providing more than 20% of aggregate network power and reactive power, most likely will require inverters that can initially operate in grid support mode, i.e., using their PLL to provide current and reactive current to augment power requirements. A number of grid-following sources can, as the black start procedure continues, switch over to voltage forming (note: to the best available information, these mode-switch inverters are not yet produced commercially).
Complicating factors like a separate communication channel for inverters and certain classes of loads cannot be avoided although power sharing between various classes of inverters can be on a ‘plug and play’ basis. A master controller will include the resynchronisation procedure. This is more difficult for BESS-inverters as, by virtue of large impedances in LV networks, there is a tendency for frequency as well as voltage to oscillate. For example, grid supporting inverter PLL being voltage-phase angle sensitive, sense oscillation in voltage phasor angle depending on total active and reactive current in the conductors at their point of connection (PoC).
As mentioned above, the isolation of the microgrid takes place at the MV/LV transformer, the latter having star-point earthing. This is essential for earth fault protection. However, inverters do not provide zero-sequence current, therefore relying on connected loads for this component of earth fault current. Differential protection can eliminate reliance on earth fault current but it requires non-standard protection equipment.
Will microgrids conquer the NEM?
It might seem a fantasy—yet, the growth of a distributed generation, principally solar, is now unstoppable. It begs the question as to what purpose the NEM with its networks of capital-intensive transmission lines might serve as from an energy aspect, distribution grids equipped with battery storage could soon become independent. Presently, there is much worrying about whether we absolutely need to retain coal-fired, synchronous generation. The worries would cease if we could build inverters, supercapacitors and large storage batteries to take over. We could well be the first in the world because we are already the standout adopters of solar!