By Phil Kreveld
Looking for ready answers for grid security in the step-change renewable transition will prove a very long, obstacle-strewn path.
Detractors of the renewable transition are justifying their objections by pointing to the loss of electrical power in the Iberian Peninsula on April 28, though the investigations are likely to be protracted as to the reasons for the breakdown. Nevertheless, the replacement of synchronous generation by voltage forming inverters as contemplated by national renewable targets may well propel Australia into millisecond decision territory in order to maintain grid security and stability.
Related article: Postcard from Spain: Blackout warning for Australia’s grid
The recently concluded election appears to have done away with the nuclear generation option, which according to its proponents would have overcome many of the obstacles associated with renewables. A balanced appraisal of the transition to renewable energy reveals impairments associated with the loss of synchronous generation, e.g:
- System strength,
- Voltage stability,
- Frequency stability and inertia, and
- Synchronising strength.
The impairments all result from the replacement of synchronous generation by asynchronous sources, mainly inverter-based resources. The retention of synchronous machines powered by coal, gas and hydro or by means of nuclear energy would diminish the above concerns.
On May 19, 2023, Australia’s energy ministers said the “time was right” to transition to a new operating model for grid security with the phasing out of the Energy Security Board, then chaired by Dr Kerry Schott, and replacing it with the Energy Advisory Panel (EAP) answerable to the state, territory and commonwealth energy ministers.
In 2021, the ESB’s alarm regarding the loss of synchronous generation in the Australian Energy Market Operator’s Integrated Systems Plan raised this statement:
“The ESB considers the existing market, and its related arrangements, are unlikely to be sufficient to ensure the commercial provision of the right mix of resources required as the market transitions towards a higher penetration of variable renewables”.
It was a prophetic statement: the Orana renewable energy zone is likely to require six synchronous condensers (synchronous generators running on no load as a motor-flywheel combinations to supply inertia and voltage support). AEMO has indicated some 45 synchronous condensers are needed in the transition to 83% renewables by 2050.
The title of this article is deliberately provocative. It stresses the uncertainty surrounding events that led to the recent separation of the Iberian peninsula’s network from France but also more generally the knowledge and experience gaps in renewable grid engineering, and makes the following points:
- Interconnected alternating current systems provide economic advantages, but with
- Considerable security risks in that faults can readily propagate, causing grid separations (islanding) and blackouts, and therefore
- Requires detection and corrective action within a hundred or so milliseconds, and inside the time scale of many grid and generator controls when renewable energy sources are utilised.
It is too soon to analyse what happened to Spain and Portugal but it is unlikely to be easily attributable to only high penetration of solar and wind generation. One could also speculate that a serious grid instability event, e.g., a very fast rate of change of frequency took place before grid control systems could react. In this regard, solar and wind generation can react much faster than synchronous generators.
The public square debate in Australia confuses energy security with grid security and this leads to confusion as to the likely problems emerging with rising penetration of solar, wind and batteries. Both energy security and grid security are essential but one does not import the other.
- Energy security treats the grid as a copper plate model (Figure 1): centres of energy consumption and sources of energy are ‘wedged’ between two electrically conducting copper plates that do not impede the flow of electrical power. The public square conversation is about there being enough energy generated to satisfy energy consumption at all times.
- Grid security is power, voltage and frequency stability, a property of the physical transmission lines, the power, voltage and frequency stability of the generators connected to it; and the predictable nature of power (not energy) consumption time patterns of load (consumption) centres.
South Australia, Ireland and Texas are the standout examples of renewable energy source penetration and therefore of emerging grid and already experienced stability challenges.
In Australia the technical aspects—the ones that relate to security in particular, are kept out of the public forum, not shared, and often glossed over in official communications, for example, from the Australian Energy Market Operator (AEMO). Note: within that organisation there is serious study of factors involved in grid security and stability. It is fair comment that in AEMO’s studies, there emerge ‘unknown unknowns’ (coined by the late George Donald Rumsfeld, USA Secretary of Defence).
The ‘unknown unknowns’, without being articulated in the electrical energy debates in the lead-up to the elections were debased to ‘fear and loathing’ of renewable energy sources or conversely in ‘coal keeper’ accusations and forecasts of nuclear holocausts. Yet, there was little or no evidence of a national effort to unravel the grid security impairments that await as the percentage of asynchronous generators heads towards 100%. An attempt is made in this article to explain the technical background to stability aspects of electrical grids based on traditional generation and to delineate the essential factors of grid stability and security because these also apply to grid forming inverters. The technical aspects apply to renewable forms of generation and this change in technologies demands rethinking of the overall control of geographically widely spread energy resources in order to retain stability and security.
The outlines of the ‘unknown unknowns’ are listed here based on electromagnetic, mechanically driven, synchronous generators being replaced by grid forming (alternate term: voltage forming) inverters: the need for synchronous condensers to support these speaks volumes about apprehension regarding grid stability and security once Australia is powered exclusively by renewable energy sources.
- Grid-forming inverters have been trialled successfully in microgrids but not in large, geographically spread networks;
- Grid-forming inverters, in general, do not behave like the electromagnetic synchronous machines, their over-capacity when short circuits occur being severely limited by comparison, but dissimilarities do not end there;
- Grid-forming inverters utilise electronics (operational amplifiers, signal filters, comparators, application firmware, etc.) of proprietary nature, making prediction of their responses to sudden load shifts, faults, loss of other generators, etc. very challenging to predict;
- The inherent property of load sharing between synchronous machines is not a native property of grid forming inverters; this is currently under investigation, there being a number of theoretical paths to follow;
- The practical experience with inverters in large grids is restricted to a sufficient aggregate capacity of synchronous generators for providing stable voltage and frequency, with additional electrical energy provided by grid following inverters, these requiring stable voltage and frequency;
- Other than control mechanisms based on grid-following inverters in combination with synchronous sources providing voltage and frequency, overall grid control and protection schemes for grid forming inverters in combination with grid following inverters exist in theory but not in practice, i.e.: renewables have been adopted into classical networks but the larger their relative power contribution, the more we head into unknown territory.
Appreciation of the challenges posed by the renewable transition requires an understanding of alternating current grids. The diagram representing the principle of operation of a ‘synchronous’ generator (Figure 2) is intended to solidify an understanding of alternating current generators. The creation of voltage and electrical current by means of magnetic induction, requires a mechanical source of power (a diesel engine, a turbine, etc,). In contrast, battery and solar powered inverters utilise stored electrical energy, and that created by conversion of photons to electrons, to produce alternating voltage and current by means of power transistor switches.
The diagram forms the basis of concepts that apply equally to all synchronous generators. The use of the plural because synchronicity implies that all generators participating in an electrical grid not only generate at the identical frequency of 50 cycles per second (Hz), but as importantly that their voltages at their terminals have closely fixed angles with reference to each other, (refer to Figure 2, and the explanation of voltage ‘phasors’). A single machine could run at any constant or variable rotational speed—in short, an electrical grid comprising a single generator supplying electrical power and energy to a single point of consumption (electrical load) is the most elemental of systems. The use of terms, synchronous and asynchronous, is a little confusing. It is therefore preferable to use the first only for electromagnetic, mechanically driven generation, and asynchronous for everything else including wind, solar and batteries.
It is impossible in a short article to sketch the many ways in which instabilities arise in alternating current networks. Therefore, one example of a generator (an inverter) losing synchronicity is illustrated in Figure 3. A very simple grid structure is shown of two generators supplying a single, dynamic electrical load, i.e. one that varies its power demand rapidly over time (this is now the case for distribution networks with a high capacity of rooftop solar systems). The two generators are in separate, renewable energy zones, a large distance from each other and from the electrical load. Assume it has been a cloudy morning when suddenly there is a break and rooftop solar suddenly supplied much more power than earlier on. The power output from the two grid forming inverters drops but generator B loses synchronicity (read the caption of Figure 3). The same could happen, in the case of synchronous machines but it is less likely because of the existence of synchronising torque, a property absent in for grid forming inverters.
Related article: It’s system strength, stupid!
Hasta la vista?
Hopefully not. One obvious path is to delay the replacement of synchronous generation with voltage forming inverters. This would also underpin voltage stability due to the ability to supply large amounts of reactive power (which must otherwise be provided by synchronous condensers).
If this task is to be performed by voltage forming inverters, they need to be energised by batteries which far exceed the required energy rating—a very uneconomical application. The reduction in inertia can be compensated by building in electronic filtering to mimic response delays that are part of the physics of synchronous generators. However, in order to assure a similar response, changes in frequency and in particular very rapid changes due to mismatching of generated and absorbed power have to be detected well within the ‘time constants’ of the electronically provided inertia.
Delving into the detail reveals nebulous practical experience—that is to say, theories a plenty but little practical experience. This is no surprise—we are in the forefront of technology. An enormous challenge, conveniently never highlighted, is the design of new control mechanisms based on the millisecond response speed of basically inertia-less generation, whether grid following or grid forming. Creating more market structures, for example for inertia, will not solve security and instability—only whole of system engineering, and a relaxation in renewable targets so as to provide adequate practical testing and evaluation periods.
