Special Feature: Off the Grid

For isolated communities microgrids are essential, writes Melbourne-based electrical engineer Phil Kreveld. Within established distribution networks they can provide increased power reliability but at considerable investment costs. Ideally, they could form part of smart grids of the future, diminishing the need for new baseload generation.

Microgrids are virtually the only way of delivering electrical energy to isolated communities—in fact, the term wouldn’t have come into popular use if it weren’t for the development and growth of solar photovoltaic (PV) and latterly of small-scale wind-generated power, holding out the promise of self-sufficiency.

As solar PV became a common sight in suburbia and regional centres and electricity tariffs continued to rise, divorce from the distribution grid grew as an idea even if some enthusiastic proponents didn’t appear to have given thought to the poles and wire infrastructure they would either have to acquire or rent.

As ‘green energy’ takes a hold, distribution companies are increasingly under pressure to permit higher power ratings for its inclusion in networks, and of other distributed generation (DG) – the technical term for not only solar PV but also for consumer-owned diesel generators, etc.

The growth in DG causes increasingly stringent grid codes to be applied by distribution companies as reverse power flow is resulting in voltage and power factor problems.

Grid codes are not uniform in the National Electricity Market (NEM), and this doesn’t help for a sensible development path for DG inclusion. The engineering reasons that underlie the application of restrictions for the inclusion of DG would also apply to the creation of microgrids within distribution systems.

Viewing the integration of microgrids in distribution networks as a means of decreasing the cost of electricity for households and businesses is not necessarily valid even though it has popular traction.

Retail tariffs are largely influenced by network charges and retail margin requirements and a decrease in energy delivery to some consumers will not decrease major cost elements of networks such as amortisation and maintenance.

If anything, it might well lift tariffs for remaining consumers not situated within microgrids as fewer consumers remain reliant on grid-delivered energy. The real value of microgrids is in enhanced energy reliability as in the event of a grid failure, they can continue to provide energy. However, this can only be achieved by taking a new approach to distribution network design.

Although the South Australian Power Network (SAPN) and AGL’s virtual power generation (VPG) project in Salisbury has some features that might be included in microgrids (e.g. cloud-based battery status information) the VPG is not separable from the network. The difference, as explained later in more detail, is the VPG, although grid-supporting (pumping battery-stored energy into the grid, when it is strained by high demand), it is not grid-forming, the latter requiring the supply of narrowly set, upper and lower limits for voltage and frequency.

Without grid-forming capacity there can be no microgrid. Ipso facto, some form of synchronous generation is part of the microgrids in Western Australia. Without grid-forming, projects such as the VPG mentioned above, although of undisputed value, represent ‘negative loads’, i.e. they can be net power suppliers but when islanded because of a grid failure, or drop in voltage, or frequency, they cannot continue to function.

The technical aspects for integrating microgrids include:

  • Detailed impedance database for distribution networks
  • Network topology studies to identify potential microgrids
  • Fault protection within microgrids
  • Review of islanding conditions, i.e. under ‘normal’ and network fault conditions
  • Reconnection to the grid

Grid-forming and resynchronisation

The easiest way of grid-forming is to use synchronous generators, however driven, diesel and gas being suitable for urban and regional locations. Wind power and solar PV can function in grid-forming and literature references included provide detailed information. However, in general, both voltage and frequency, but more so frequency, have wider band limits than is the case for synchronous generation.

This is usually not a problem for the microgrid in its isolated operation mode, but it is more of a problem when resynchronisation has to take place on reconnection to the grid. In instances where synchronous generation is employed, but where it represents a small fraction of the total microgrid demand, resynchronisation is also more challenging.

As is the case for load-sharing between synchronous generators, droop control is also central to the task for the operation of microgrids. Frequency droop with increased effective power demand and voltage droop with increased reactive power demand, form the control characteristics for synchronous generators connected in a high-voltage (HV) system with typically high inductive reactance (X) to low resistance (R) ratios.

Known also as stiff networks, they stand in contrast to medium-voltage (MV) and low-voltage (LV) distribution, which typically has X/R ratios of five, and less (weak or slack networks), and at network fringes is essentially resistive.

As shown in the paper by Laaksonen et al1, the droop characteristics for weak networks (R>X) differ so droop of frequency and reactive power demand, and voltage and effective power demand, provide power-sharing and grid-forming. Depending on X/R ratios, droop characteristics can have the same form as for stiff networks but using park-transformed voltage and current parameters2.

The grid-forming process in the absence of participation by synchronous sources can be subject to voltage and frequency instability through the lack of rotational inertia. However, by not operating wind and solar sources at maximum power, a form of synthetic inertia can be effected. This is achieved by means of torque control in wind generators, and by not operating solar PV generation with maximum power tracking.

Islanding and protection

Islanding conditions also need to be revised. As matters stand, grid codes require islanding (with grid-controlled reconnection) in the cases of grid over and under voltage, as well frequency instability. None of the grid codes allows for fault ride-through, and this part of the codes would require revision. Islanding conditions, ideally, would then allow microgrids within a distribution network to operate independently when fault ride through conditions have timed out, thus providing superior reliability for consumers.

As already indicated, subsequent reconnection requires resynchronisation3, and particularly for non-synchronous energy sources, this can be a time-consuming process unsuitable for practical situations.

The advent of increased distributed generation gives rise to changes in network protection. For example, directional power flow relays will not detect reverse power flow. Impedance relays are likely to locate a network earth fault, further downstream from the relay than is actually the case by virtue of reduced fault current brought about by reverse power flow from DG. In the case of inverters (wind or solar), there is generally insufficient current to feed a fault therefore requiring new approaches4, for example, differential relaying.

Challenges in distribution

Distributed generation broadly encompassing microgrids, requires detailed topology and impedance information. The case for this has been adequately made by Alexandra von Meier et al5 on the application of micro-synchrophasors in the study of power flow in LV distribution. The use of global positioning system GPS-based phasor monitoring in HV transmission is well established and this is now being translated to very small-phase angle differences (micro phasors) as seen in LV networks.

Gathering of near real-time power flow data using supervisory and control data acquisition systems (SCADA) with GPS time-stamped, micro-synchrophasor data can support smart network design. The AS/NZS 4777.2 standard specifying demand response enabling devices (DRED) for distributed generation is a harbinger, it could be argued, for smart networks – admittedly a broad term – but one the distribution sector needs to factor in when re-evaluating the suitability of current networks for meeting reliability.


  1. Hannu Laaksonen, Pekka Saari, and Risto Komulainen, Voltage and Frequency Control of Inverter Based Weak LV Network Microgrid
  2. K De Brabandere, B Bolsens et al, A Voltage and Frequency Droop Control Method for Parallel Inverters; 2004 35th Annual IEEE Power Electronics Specalists Conference, Aachen, Germany. pp2501-2507
  3. Di Shi, Senior Member, IEEE, Xi Chen, Senior Member, et al, A Distributed Cooperative Control Framework for Synchronized Reconnection of a Multi-Bus Microgrid
  4. Manjula Dewadasa, Protection of Distributed Generation Interfaced Networks, Thesis, Doctor of Philosophy, Queensland University of Technology, July , 2010
  5. Micro-Synchrophasors for Distribution Systems Alexandra von Meier, David Culler, Alex McEachern