Betting on batteries.

Intermittent generation, storage and two-way grids. If you think that sounds complicated, you’re probably right.

I attended an industry event last week and it was clear that from an electricity perspective that we are intending to go full steam ahead building out more wind and solar.

The obvious issue with this is how do you resolve the grid stability and 24/7 demand requirements with increased intermittency?

The following article is a quick overview of this issue. This is a highly technical subject and even some of the market mechanisms required to make this work will be very difficult to explain let alone overcome. I haven’t even tried to understand how this would be coordinated between customers who are also potentially generators trading arbitrage, lines companies, the system operator and the commercial generators. Any one of the topics here is worthy of a standalone article so I’ll do my best to provide a basic overview and some of the things that need to be considered.

Batteries, wet and dry.

The big four generators approach to this issue appears to be twofold.

1. Repurpose the hydro schemes over time to act as firming for wind.

2. Use BESS (Battery Energy Storage System) coupled with solar to give solar a generation time preference.

I have already written on the re-purposing of hydro schemes in “hydro peaking”, so I won’t spend too much time on this other than to briefly say.

• It makes sense in a dry year to preserve the hydro schemes.

• Batteries are not suitable for this application as they will never be able to scale sufficiently to support wind generation. Wind as discussed in “market guardrails” has a surprisingly unform generation output across the length and breadth of the country. The wind farms all basically generate at the same time, as such it is feast or famine. With 1400MW of wind generation currently installed on the grid it’s far too much for batteries to support. This is also why the wind farms have largely been consolidated by the big gentailers with hydro in their portfolios. They are essentially the only ones that can make it work commercially.

• The key issue I see with this scenario is that it’s counterproductive much of the time. Fundamentally it does not generate more power. It generates the same amount of power bi-modally, alternating between wind or hydro. This significantly increases the asset base that generates the power and by implication the per unit cost of electricity. Apart from particularly dry years I expect this to increase the wholesale price of electricity due to the underutilization of the lowest long run marginal cost of generation which is hydro.

EA – Future Operation of New Zealand’s Power System

The Electricity Authority (EA) is giving the intermittency issue much consideration too and last year issued a consultation paper called “The future operation of New Zealand’s power system. Consultation Paper”, and more recent a follow up that is based on the feedback from the first consultation.

Note, as I had observed in “market guardrails” we have boxed ourselves into lofty goals of 98% renewable by 2030 which limits our options and comes with some really significant technical problems, not least of which is how to pay for it.

A few key extracts from the EA consultation papers that sum up their views on the issue.

The flow of electricity in New Zealand’s power system is largely uni-directional at present. Typically, electricity flows from large power plants, located near fuel sources, to consumers via transmission and distribution networks.

Transmission-connected power plants predominantly comprise conventional synchronous machines powered by hydro, thermal (coal and natural gas), and geothermal fuel resources. In 2022, synchronous generation accounted for approximately 93.5% of total electricity generated in New Zealand.

However, technology, and the need to reduce greenhouse gas emissions, are driving a move towards the use of inverter-based resources (IBRs) for generating electricity. An IBR is equipment that uses an inverter when functioning. Examples include:

(a) wind generation

(b) solar photovoltaic generation – ranging from small residential rooftop installations to large (utility) scale solar farms

(c) a battery energy storage system – ranging from a small residential wall-mounted battery and an electric vehicle battery to a large (utility) scale battery.

Boston Consulting Group (BCG) estimates 4.8GW of new large-scale renewable generation will be needed to achieve 98% renewable generation by 2030.55 Much of this new generation will be IBR generation. There is also significant growth expected in small-scale residential IBR generation, especially rooftop photovoltaic solar generation and electric vehicles with batteries that can inject electricity into the local electricity network.

Ok, so there is some nuance here we need to explain.

The uni-directional grid, as we have historically known it, was simple. It was comprised of a potential energy source at one end in the form of water, gas, coal, or steam. Generation machines, known as alternators, convert this potential energy into electricity, which then flows through the network to the end users.

The key feature of this system is the ability to modulate the amount of potential energy required to match the consumer demand, typically with a valve or governor. It is easy to visualize by thinking of it as being similar to controlling a cars speed by modulating the fuel supply to the engine using the accelerator.

The grid of the future as the EA see it is bi-directional and will be comprised of more generation sources that are not able to be modulated to match demand. As a result, they see the need for storage in the form of batteries to absorb the excess generation and then discharge it into the grid during periods of insufficient generation.

This overview may sound simple but to run a bi-direction grid that is stable and reliable is a big challenge.

Betting on batteries

The EA and many others in the electricity sector are betting big on batteries.

The expectation is that hydro lake storage, BESS grid scale batteries, electric car batteries and home battery systems will combine to create a bi-directional grid.

As far as I can gauge the vision is that wind will be the primary generation source which will be buffered by water in the hydro schemes. Lake levels will be kept high and even spilled to provide storage for low wind periods.

Solar will be discharged to batteries behind the meter (between the solar panels and the grid connection) and grid connected batteries will be charged during high generation periods and discharged during low generation periods to create a grid that can meet demand 24/7 independent of the weather. Batteries being key to decoupling weather driven intermittency from the requirement for a constant supply.

Of course there is also a push for more geothermal which is very good and important for reasons I will touch on later.

Assumptions

This plan is underpinned by some key assumptions as follows:

• The uptake of EV’s will be high and that they will be charged mostly during the day.

• Battery technology will scale and be cheap.

• Solar and wind hardware will continue to get cheaper.

• Domestic solar and battery systems will widely taken up.

• Capital will be cheap.

• Complexity can be managed.

• Grid stability can be maintained.

Testing the assumptions

EV uptake:

NZ’s EV uptake peaked in 2023 when the subsidies were removed. 2024 sales were down from 2023 and are not expected to recover in 2025. Is this trend baked in? In the absence of subsidies probably, however an oil shock would reverse this for sure. Watch Iran closely.

The EV is often a second car and used around town. This means than charging during the day while solar generation is at a peak is not particularly practical as most of them will be parked in office car parks and not at home in the shed connected to the grid.

Battery technology will scale and be cheap:

There is a lot happening in the battery space, and there needs to be. In grid scale terms the levels of storage that the current technology can provide is tiny. We are talking about MWh systems when multi GWh systems are required. For example 1 GWh of battery storage would supply about 1/3 of the North Island load for 1 hour at the time of writing this.

There is a fundamental relationship between material quality, scarcity, and functional output. High quality materials are structurally highly ordered and pure. This makes them rare. But most importantly it makes them low entropy. Low entropy is the key to their performance as a storage device.

A lot of people are very excited about new battery technologies such as sodium-ion, iron-air, zinc-ion, aluminium-ion, and possibly solid state or organic carbon batteries.

In this regard I’m in the thermodynamic realities camp that suspects hoping for a cheaper material to do the same job is like hoping stone will replace tool steel. That said the lithium prices dropped by a whopping 69% last year which if sustained could see the high performance lithium-ion systems become cheaper.

The jury is still out on the ability of battery technology to meet grid scale storage requirements, I’m sceptical and I’ll be watching with interest.

Solar and wind hardware will continue to get cheaper:

We are betting on a one horse race here. The fate of this plan largely resides in the hands of China’s political class and their market aspirations. As we have seen in many markets if China wishes to monopolize they flood the market and destroy the competition. They have done this successfully with rare earth metals processing, solar panels, EV batteries and increasingly EV’s themselves.

I would suggest that the low cost of solar panels for example has much more to do with oversupply in the market than technological advancements and scales of economy. Solar panel supply hugely exceeds demand and this race to the bottom has resulted in China consolidating over 80% of global manufacturing.

Indications are that China is making moves to re-balance supply and demand now that they have the market cornered. I think that we may have seen solar panel costs at the lowest levels they are going to get and they could possibly increase in the next couple of years as supply and demand rebalance. However, polysilicon remains the dominant material, and this is very energy intensive to produce. Coal is currently cheap, and China now uses ~58% of the global supply so I could be wrong on this one if coal price remain low.

Wind technology was seeing a price decline up until around 2021 which saw this trend reverse due to supply chain disruptions and increases in commodity prices of materials. Prices have since stabilized at the higher level. Looking ahead, due to the high material intensity of wind turbine manufacturing the cost ironically is closely tethered to the price of coal, oil and gas as major energy inputs to the material supply chain. If I were a betting man, I don’t see a return of the declining cost trend.

Domestic solar and battery systems will be widely taken up:

There is definitely potential for more homes and farms to have a solar array. There may never be a better time than now to do it given the global over supply of panels out of China.

A home battery system however is a different prospect. It is much harder to get a return on investment with this system which still requires a high capex investment.

One of the great ironies of this plan is that there has to be a commercial incentive for arbitrage at a household level. People need to be able to control if and when their batteries would re-inject to the grid.

I can’t circle the square on how we are expected to believe that NZ’s electricity prices will reduce if we have lower utilisation of the transmission network, because more people are generating their own power from less efficient systems while receiving a premium for the power they supply?

As we see in other jurisdictions domestic solar with contracted grid supply rates cannibalizes grid scale solar, which means we need more batteries.

Capital will be cheap:

Predicting interest rates is outside my wheelhouse but I will say that global financial markets are highly leveraged.

Probably a more important consideration is that New Zealanders homeowners are generally house rich and cash flow poor. Not a lot of people can afford to buy a solar system outright and even less can add a battery to it. Wider adoption will be almost entirely be contingent on the finance terms available.

This extends to industrial scale wind and solar systems too. These projects are not huge a huge cost, I’m thinking hundreds of millions as opposed to billions, but still expensive. New entrants need to attract finance and are very vulnerable to the cost of capital.

Lifecycle is a further capital consideration. Battery systems currently have a 10-15 year lifespan or ~4000-6000 cycles. Solar panels are a bit longer in the 15-20 year range but similar. These are short lifecycles in energy system terms and this could be a major challenge if the assumption that these technologies will get cheaper does not play out.

Complexity can be managed:

A bi-directional grid with decentralized generation is far more complex than a uni-directional system. Complexity involves a law of diminishing returns. This is because as complexity increases so too does the energy required to service the complexity. This system will require more coordination and be far more unpredictable, as such it needs a lot more systems to manage.

Submitters to the EA consultation paper stated “this would likely require significant investment in automation, communications with smart meters, real-time systems, managing big data, and the use of data analytics to provide better visibility of the distribution network and its connected distributed small-scale generation”.

I don’t think the implications of increasing complexity with more failure modes should be underestimated.

Grid stability can be maintained:

This is a tricky one which again introduces further complexity.

The grid of old managed voltage and frequency control by having synchronous large rotating generators directly coupled to the grid and all generating at 50hz. In addition to being able to be modulated in response to load, as described earlier in this piece, they also provided inertia to stabilise the grid frequency, which must be maintained within very tight tolerances.

The proposed bi-directional grid depends on large amounts of asynchronous generation that is not directly connected to the grid. It is instead connected by electronics such as inverters and rectifiers. These systems do not provide any mechanical inertia.

To try and explain this to the widest possible audience I’ll use a similar analogy.

If you have ever ridden a bicycle without a freewheeling hub you will appreciate this. You can’t stop pedaling because the bicycles inertia is mechanically coupled to the pedals. The new grid is more like a freewheeling hub the bikes inertia is decoupled from the pedals by the freewheeling hub.

Could batteries provide a grid forming service. Yes, to some extent, but with caveats.

Firstly, to achieve this at a grid level we would need a huge amount of battery capacity. Secondly it would require grid stabilization as a service. What I mean by this is that it would need battery systems reserved specifically for the purpose of grid stabilisation, not for the charge low sell high arbitrage that pretty much all battery business cases rely on.

Batteries are fast response and can provide voltage up, voltage down services. They can also provide grid frequency support with inverters specifically designed for this purpose. However, flat batteries can’t support anything which is why we would need a market for grid support as a service. Batteries providing this service would need to stay charged and available to respond. They could not be cycled outside of a narrow band in the same way that batteries intended for arbitrage would be fully discharged. We would need to pay the operators to keep them charged.

It is also important to note that in the North Island we are losing a lot of generation with associated rotating mass from the grid. The 377MW Stratford combined cycle gas turbine will be retired and we will also lose about 70MW of co gen from Fonterra’s Whareroa plant as the steam system is electrified next year. There is about another 100MW of geothermal coming online next year which helps but the overall amount of inertia in the system is decreasing.

Wide boundary analysis:

As I have noted in New Zealand Energy Strategy our energy system is at it’s core a physics problem.

Given that the economy is in essence an open thermodynamic system, that requires an increasing energy surplus in order to grow, EROI is of critical importance.

The pathway laid out by industry leaders and the EA relies on lower EROI systems that need to be built with excess capacity and have storage bolted on. The more support and backup systems that are needed to make this work the lower the EROI gets.

With relatively short effective lifecycles, much of the system hardware will need to be replaced every 15 or so years which further reduces the full lifecycle EROI.

As stated in “energy debt trap”, just adding generation is not enough. It needs to be energy that provides high levels of surplus. The reason I say “surplus” energy is because we need energy that is available to the productive economy. Not energy that is consumed in the circular pursuit of energy.

I fear that this Rube Goldberg machine we are planning on creating is just that.

Some food for thought folks and I stand to be wrong on much of this. I think one of the key takeaways here is that it is all contingent on technology getting better and cheaper. There are strong arguments that it could go either way.

P.s. I’m travelling at the moment so the posts and replies may be a bit sporadic over the next couple of weeks.

Have a great weekend.

Larry