By Phil Kreveld
The report into the Iberian Peninsula blackout affecting Spain, Portugal and Morocco, rather than giving ground for kicking the slats out from underneath renewables, makes the case for dispassionate analysis and for appraisal of engineering-based solutions that address renewable energy source limitations. Fact is that the public forum debates are either concentrated on these limitations or a complete denial of their existence. Both sides render a disservice to a proper growth path for renewable energy sources.
Wading through the heavily redacted report of the Iberian grid failure on 28 April, the first conclusion is that it is a chronological account which doesn’t shed much light as to how it might have been avoided. It is perhaps an unrealistic expectation given the level of redaction, in anticipation of legal repercussions. Yet, there is one aspect that emerges, deserving our attention: reactive power. And it receives a fair amount of emphasis in the official Spanish language report. Anti-renewable folk have already concluded their case that we are heading the wrong way. Others think the lack of inertia is to blame. Yet close reading of the Spanish report in the native tongue (to avoid translation machine errors) reveals speculation that the nexus between reactive power and voltage was a very important factor in the blackout.
Related article: Postcard from Spain: Blackout warning for Australia’s grid
The essential nature of reactive power, and its function in voltage control in an electricity grid doesn’t receive the same attention as active power. It is understandable because only active power provides energy. Most energy consumption requires in addition to the essential active power, a component of reactive power. Motors, in particular, draw reactive power, required to maintain their rotating magnetic fields. Incandescent lighting and resistor-based heating do not ‘draw’ reactive power. Electricity 101 students learn this in year one; motors require ‘lagging’ current (see Fig. 1), so do fluoro ballasts and transformer-based lighting like LEDs. Electrical engineers refer to these electrical loads as ‘absorbing’ reactive power. In actual fact, there is nothing of the sort happening as reactive power bounces back and forth between generators and their electrical loads. However, we’ll stick to the convention and therefore conclude that if loads absorb reactive power, then there must be generators providing them reactive power. Note that reactive power has the same units as active power (watts) but in order to distinguish it, the units are called volt-amps reactive, shortened to ‘var’.
The Spanish report stresses the lack of generated reactive power in the Iberian grid following its disconnection from France. Sources consulted other than the official report describe the link to France as weak. In Fig. 2, this is explained graphically. Note that the power-voltage curves exhibit a maximum active power point whereafter attempts to transmit more power results in voltage collapse. In this context a weak link operates near the ‘nose’ or maximum power and permissible voltage. There are a series of curves, and note that some permit the same amount of power to be transmitted with much higher voltage at the power receiving end of the line. The latter relate to loads absorbing less reactive power, or in electrical language, they have a higher power factor (see Fig. 3 for a further explanation).
S=√(P^2+Q^2 ) and power factor is P/S and S is apparent power
Some of the explanations in the Spanish report are convoluted and require a detailed line diagram of the Iberian grid including Portugal and Morocco. However, let’s not avoid a significant matter relating to renewables providing reactive power support mentioned in the report. Please refer to Fig. 4, which is circle diagram. In essence, both solar and wind generation operate according to the principles as shown in the diagram, which explains that reactive power can be supplied but at the cost of a reduction in active power. Note that in this regard, synchronous generators can supply their rated power without limiting reactive power to the same extent as wind and solar generation. On average the ratio of reactive power to active power can be three to four times active power for synchronous generators whereas 15% of active power is typical for solar and wind.
In Fig. 5 the elements of power transfer from a generator of whatever type to a usual load, one absorbing reactive power, is shown, i.e., as indicated by the black triangle. The transmission line connecting the generator to the load also absorbs reactive power (this is an inherent property of lines of less than 100km—longer lengths provide a mixture of absorbing and sources of reactive power). Clearly, the generator in the diagram, in addition to supplying active power (for simplicity’s sake active power loss in the transmission line is neglected) for the load also has to provide additional reactive power for the transmission line. There are devices for reducing both the reactive power demand of load and line, termed compensation devices, a common one being the var compensator, a static device, and the synchronous condenser. The latter is a synchronous machine running as a motor providing reactive power but drawing very little active power as it does not need to provide mechanical power.
Related article: Milliseconds for decisions—or is it hasta la vista?
Spain attributes a significant part of grid collapse to voltage regulation problems; insufficient reactive power supplied by solar and wind to reduce outside-limits high voltage at distant buses (grid connection points). The Government report indicates that low active power transmission in distant connection points caused voltage rise (we won’t go into details for the sake brevity) caused by the Ferranti effect. The important thing to think about also for Australia is that our grid also runs virtually entirely with grid-following inverters and that includes battery storage. Stable voltage is therefore essential. Now for some quick maths based on Fig. 3. If by 2050, our power demand is 300GW (Australian Energy Market Operator integrated systems plan) and the transmission plus consumption of active power factor is 0.9, then 145 gigavars have to be generated in reactive power. Commercially this has to be supplied by generators, and referring to Fig. 4, at the cost of diminished active power (and revenue) or by massive investment in reactive power provision such as synchronous condensers.
Our lesson from this is to be a Sancho Panza to Don Quichotte, the Spanish knight errant, fond of tilting at windmills, and to reflect sensibly, like the valiant but eminently sensible sidekick to the ‘Don’ that planning ahead really pays off. Voltage regulation is a major item in an essentially inverter-driven grid, come the renewable targets closing in. And as we will discover, it needs to be controlled on a whole-of-grid basis, designed not by committees, with interminable meetings and delegates keen on defending their own patch, but a central grid stability authority with the very best of power engineering talent.
