Virtual inertia in practice: How South Australia’s second big battery made its mark

Grid forming inverters + battery = System strength

Recently Hitachi ABB Power Grids held a great webinar showcasing their Virtual Synchronous Machine [VSM] and discussing its use with South Australia’s second big battery – the installation at Dalrymple, also known under the acronym ESCRI, which is a 30 MW/20 minute battery.

In this note I provide a financial analyst’s perspective on what was showcased by Hitachi ABB Power Grids. As a financial analyst I need to make a judgement about what works and doesn’t work in order to understand the opportunities and risks of  decarbonizing the grid.

The risk to me is always that not having the power engineering background I don’t understand the limitations or other factors of something technical. But this is standard investment bank research risk where new technology is always being evaluated for its commercial significance. The mitigation is to cross check and do your due diligence.

Inertia – resistance to change…

Inertia provided by rotating machines (coal generators) was essential to controlling the 20^th century grid. If alternatives do not exist, then decarbonsiation and closure of coal stations is delayed.

If, on the other hand, we can accept that grid forming inverters can actually do a better job than the old system by making grid control more autonomous, more resilient by being able to island individual segments, and more cooperative, then we can be more confident of predicting closure of coal stations over the next ten years. It goes further though, because it points to the pivotal role of batteries and software.

It points to new markets in transactive energy and it points to a terrific device market investment opportunity. SolarEdge the world’s largest manufacturer of inverters,  only started 15 years ago.

It’s an Israeli company with a smaller domestic market than Australia. Businesses such as Hitachi ABB Power Grids that have taken the risk to invest in the future early on will, if they deliver the products at the right price points, end up earning great returns as they will have the solutions while other manufacturers are still trying to hang on to the old world.

In Australia, the Australian Energy Market Operator is going to have to start making the same decisions. The delay is already costing consumers and investors. The coal generators are going to close, physical inertia is going to fall. System strength issues are arising.

Australia has a very thin transmission grid. These new systems offer a way for Australia to keep its thin grid but still have excellent system strength and to build on our already commanding position in microgrids and behind the meter generation. This is a great time for skilled engineers and business people to help Australians regain the leading edge in energy economics.

Dalrymple Bay proof of concept

Dalrymple/ESCRI is proof of concept of virtual inertia  and represents the  world’s largest, albeit still small, grid connected micro/mini  grid  mostly acting as part of the main grid in South Australia, but able to seamlessly disconnect and operate as an island with up to 8MW of load, even when the nearby 90 MW Wattle Point wind farm is running. Of course the wind farm has to spill much of its output when this disconnection occurs but the lights stay on.

Figure 2. Source: Hitachi ABB Power Grids

Dalrymple has also already provided significant other benefits to South Australia generally, contributing to a reduction in the amount of rotating machine inertia required and helping South Australia to keep the lights on when the Heywood interconnector was disconnected.

A full discussion of Dalrymple/ESCRI can be found at Dalrymple knowledge sharing portal. I guess I have to add that I think this will go down as one of the best investments ever by ARENA.

The note draws heavily on the  Hitachi ABB Power Grids webinar presentation.

A stable grid needs frequency control, voltage control, fault ride through and fault recovery

“System strength is the ability of the power system to maintain the voltage waveform at any given location with or without a disturbance”.

The range of services and where various products fit in was portrayed as follows:

Figure 3. Source: Hitachi ABB Power Grids

As portrayed in the diagram a “Grid forming BESS”  contributes to all four quadrants of the services circle. BESS = battery energy storage system.

What is inertia?

Hitachi ABB Power Grids defines inertia for this purpose:

“Capability of the power system to resist changes in frequency by means of ..  …equipment that is electro-magnetically coupled  with the power system and synchronised to the frequency of the power system.”

Inertia provides the momentary power that gives time to generators to respond to load changes, or generator output changes across the grid. It provides time for other system security devices to operate and provides the buffer essential to managing frequency and voltage over the grid.

A major concern raised by AEMO and many others is how to control the grid as coal generators with their huge rotating masses are closed reducing system inertia. The reduction in inertia contributes to system strength weaknesses, impacts on the grid frequency stability and can have other implications for voltage.

Several examples of this can be seen in the NEM already. But of course give engineers an engineering challenge and a solution is soon to hand. Arguably grid forming inverters and batteries will eventually lead to a more resilient, more capable grid, and that the future will rush in at a great pace once the naturally safety first engineering industry becomes happy that the technology can do the job. As usual the technology will be years and years in front of the rules and regulations appropriate for the “new order”.

What is a grid forming inverter?

I already discussed this in a previous article for RenewEconomy but that covered a bunch of other stuff as well. The most helpful slide I’ve seen is from Bob Lasseter at the Wisconsin Energy Institute.

P&Q refer to active and reactive power and droop control adjusts active power of the asset to control frequency.

Essentially most inverters in operation in Australia today, including those found at most wind and solar farms, require a voltage reference to operate and are therefore grid following. A grid forming inverter provides its own voltage source. The battery provides the power, voltage and current to operationalise the Grid forming inverter.

There is a very interesting question as to how millions of these grid following inverters all interact and whether they can be made to do “more”  and power engineering researchers are already beavering away on that.

At the expense of a small detour, “transactive energy” is proposed as a way forward. So there is both  a “division mode” where the micro grid can seamlessly island and reintegrate, as happens at Dalrymple, and there is a “unification mode” where the systems adapts to changing conditions and incidental disturbances by the cooperation and resource sharing of the participating microgrids.

Scaled up enough and we have the full grid. Because each “section” can island the system is more resilient and because resources are shared it could end up lower cost.

Figure 6. Shahidepour “transactive energy shahidepour pdf

But lets get back to the virtual synchronous machine [VSM]

What is the difference between  inertia and fast frequency response [FFR]?

According to Hitachi ABB Power Grids:

FFR is a response in 2 seconds or less but does not have a single definition globally. (I would note that batteries can provide FFR in the order of say 300 milliseconds, and my belief is that this would not look all that different to inertia from the system’s perspective at least in terms of managing frequency) It is important to understand the distinction between FFR and virtual inertia though so the limitations of one don’t prevent the use of the other where it can actually do the job. And in fact as well as all the great diagrams and the power transfer equation this was the main learning from the webinar.

“FFR isn’t an inherent response

FFR requires measurement, detection, processing, filtering and activation –inertia does not

FFR is an important service to compensate for less inertia in the power system but is does not replace inertia

FFR can falsely trigger or fail to trigger when attempting to measure frequency and RoCoFvery quickly after a major power system fault

FFR can be provided by grid following and grid forming inverters; inertia only by grid forming inverters”

The  virtual inertia is provided by the grid forming inverter because it provides a voltage source and injects or receives power automatically when the grid voltage moves away from the battery voltage. The following power transfer equation presented at the Hitachi ABB Power Grids webinar states that power transfer is approximately equal to the phase angle of the two voltage sources, thus provided automatic, inertial, “natural” corrective force. Or at least that is the way I read it as a financial analyst.

Figure 7. Source: Hitachi ABB Power Grids

Any change in frequency automatically (by virtue of the power transfer equation)  causes an injection or absorbence of power which acts exactly as if an equivalent amount of rotating machine inertia was available.

Again I would note, as a financial analyst, that the virtual inertia is only a secondary service of grid forming inverters, the main attraction is their ability to operate independently of a  main grid voltage reference.

Being a digital box much like a guitar players amp emulation, all the parameters are “tweakable”.

Compared to synchronous condensors

This slide from the Hitachi ABB Power Grids presentation shows some economics based on 2019 revenues.

Figure 8. Source: Hitachi ABB Power Grids

However, it’s important to look forward to the 2021 and beyond revenue opportunities. We note the Australian Energy Markets Commission is looking into a rule request for a fast frequency market.

More broadly if you accept that the whole control scheme for the grid is going to change, you will end up being cautious on predicting future revenues based on out of date market designs. But you will be confident that your product will have a market of some kind.

I’d also add that the battery you add in may have other value stack opportunities including avoiding “cap” costs and energy arbitrage. These values are small now but will grow in the future as the amount of dispatchable generation falls relative to demand.

Also, battery costs per unit of energy stored fall significantly as duration increases. Right now that cost fall is less than the revenue fall off  and so battery durations are short but again I expect that to change. Indeed we can already see a tendency around the world and even in Australia for battery durations to increase, but that again is really a separate story.