Wholesale Demand Response in the NEM: What’s it all about?

For energy market enthusiasts, October 2021 is a date in the calendar worth circling. We’ll see the introduction of two significant rule changes into the National Electricity Market (NEM) on Australia’s east coast.

October 1 sees the start of 5-minute settlement (5MS), and from October 24 energy consumers will be able to sell demand response into the NEM for the first time under the new Wholesale Demand Response Mechanism (WDRM).

We’ll talk more about 5MS over the coming weeks I’m sure, but the focus of this post is Wholesale Demand Response (WDR).

Demand Response: the basics

In a generic sense Demand Response (DR) is a form of flexibility service that involves energy users – that is, the demand side of the market, as distinct from generation – turning up, turning down, or shifting demand in real-time in response to market and network conditions.

In Australia in the NEM it’s more typically associated with larger energy consumers reducing their electricity usage during times when the balance of demand and supply is tight, and therefore price on the spot market is high. Simply put, sometimes it’s more cost-effective to reduce grid demand than to pay for that extra supply. In Western Australian WEM, with its capacity market design, demand response means registering demand side flexibility for capacity credits (to earn revenue), or flexing load during peak trading intervals (to reduce cost).

Almost every market around the world has some way for consumers to create value through demand flexibility, but the devil is in the detail.

So DR isn’t a new concept in the NEM, but up until now the only way for a customer to derive value from offering it has been via a contract with their energy retailer. This is because it’s typically the retailer who has exposure to the wholesale market price and so retailers can negotiate contracts with their customers that limit their exposure to high market price events through the customer simply using less grid-supplied energy during those same events.

The customer can achieve this reduction in grid-supplied energy however they like, but typically it’s been through either turning on embedded generation (historically diesel gensets, in the future increasingly though other DER such as battery storage) so that their overall energy consumption doesn’t change but the amount of energy they’re drawing from the grid reduces, or by actually reducing energy consumption on site, perhaps by turning down a pot line, turning off chiller plant, stopping pumps and so on.

So, if DR has been around for a while, then what’s changing? Well, what’s significant about the new Wholesale Demand Response market is that, for the first time, energy users will be able to sell demand response directly in the wholesale market independent of their retailer, in a very similar fashion to the way existing generators sell energy into the market. Energy consumers will either be able to participate directly, or via specialised aggregators called Demand Response Service Providers (DRSPs). The DRSP is a new role created to support WDR.

This presents yet another threat to the existing business model of the energy retailers. Already being buffeted by the winds of change associated with the global shift to more distributed and decarbonised energy systems, they now face the prospect of competitors hitting up ‘their’ customers (you’ll often hear utilities, retailers and network alike, talk about how a customer is theirs…) to offer them a WDR deal.

Whilst this new service will create a more open and competitive market for energy consumers wishing to generate some additional revenue from the energy assets, it’s not without its complexity. A couple of the most significant questions for businesses looking to participate revolve around:

1. Eligibility – will their site or sites be allowed to participate and how is this assessed?
2. Commercials – How much might participation be worth and how might participation in WDR impact the broader commercials of their energy assets, remembering that there are multiple value streams associated with ownership of DER and chasing one can often come at the cost of another.

Eligibility

The new WDR market has a number of qualification requirements that mean not every energy user with the ability to reduce their demand will be able to participate. Here’s a summary of some of the main points:

• You need to be a ‘large’ customer as defined by the State-specific classifications. Typically that means your site needs to consume more than 100MWh or 160MWh of energy per year (varies by state)
• A minimum ‘dispatchable unit’ size of 1MW for an individual NMI or aggregation of NMIs. For an aggregation there’s no minimum size on the individual NMIs (but they must all be ‘large’ sites as per the point above)
• The load needs to be controllable for the purposes of scheduling, so walking down to the switch room and turning something on or off by hand won’t cut the mustard.
• It seems like the distribution network owner has power of veto if it considers that by customers participating in WDR they might pose a risk to the network.
• The load needs to be predictable for the purposes of baselining.

This last point is perhaps the most significant as it introduces two important and related concepts, predictability and baselines.

AEMO require that loads be predictable, both so that they have a good sense of what DR response can be offered into the market and also to assist in settlement following a DR event. Unlike energy generation, because DR occurs in the context of a site’s background energy use it’s tricky to measure directly. The solution is often to try and estimate it by comparing actual site consumption with an estimation of what would have occurred absent of the DR. The difference is the response itself. There are a few approaches for generating this estimate, of which baselining – using the history of a site’s demand – is the most common.

In AEMO’s own words, a baseline is the counterfactual energy amount for the participant that is dispatched for demand response. Generally, baselines represent an estimate of the consumption per trading interval during a day, based on a history of like days in the near past.

AEMO is proposing a baselining mechanism, the CAISO “10 of 10” framework, that’s borrowed from other DR programs around the world and works by selecting a number of recent ‘qualifying’ days for a site and then averaging the energy use, interval by interval, across those days to provide the baseline load shape. When a site is being evaluated for qualification into the WDR market then energy consumption during the relevant period is compared against the baseline and must sit within some published error margins. There are a few other wrinkles, including some adjustment to the baseline based on energy use in the immediate hours leading up to the assessment window, but that’s the gist of it.

You can read up on the detail of the baselining approach here from AEMO’s docs.

Practically speaking, this means that if a site’s energy consumption is not sufficiently consistent across days and weeks, and there are plenty of reasons why it might not be, not least because of behind-the-meter solar PV, then it won’t qualify to participate in WDR. If that’s the case, there are still ways for customers to access DR revenue but it will likely be via an ‘old skool’ bilateral arrangement with their retailer.

Commercial Value

Assuming your site can qualify for WDR, how much money might you make, and will it be worth the effort?

Let’s tackle those in two parts, first up, what might WDR be worth to you? This will depend on a number of factors, including bidding behaviour, contractual arrangements (how much value will the DRSP share with a customer) and site operations (can the site respond to every opportunity), to name a few, but standing tall above these will be the actual value of the wholesale market itself, and in particular the frequency and duration of high-priced events.

Calling volatility in the NEM is a mug’s game and so I won’t attempt it, except to say that high (and low) price events will continue into the future and the shift to 5-minute settlement on October 1 will change things up a bit (possibly a lot); it’ll probably reward fast response assets, so things like battery storage, as slower responding assets won’t have the benefit of the extended settlement window to fire up.

But despite not having a crystal ball we can still look to recent market performance as a proxy for what WDR might look like in the future, so let’s do that. Here we’ve had a go at assessing the value that 1MW of demand response may have captured in each NEM region over the past 24 months (keen NEM observers note this doesn’t yet include the data from May 23rd onwards i.e. Callide Power Station craziness). We looked at four threshold prices ($300, $1000, $3000, and $10,000 per MWh) so that’s the price that the market would need to hit before the demand response was required. For simplicity’s sake, we’ve assumed 100% of the value would go to our customer. In reality, energy users offering WDR via a third-party aggregator will have to share the spoils with the DRSP.

Finally, we also considered the number of events, or the ‘effort factor’ required to earn that revenue. Obviously, lower price thresholds occur more often, giving you lots of opportunities to earn small amounts of money, but also require way more participation. It’s interesting to think about what a sensible trigger price might be that maximises commercial return whilst minimising effort (hours required to deliver the service), particularly given any DER can often make (or save) money for the asset owner from a number of other value streams.

My takeaway from the data is that over the last 24 months of wholesale data the ‘sweet spot’ for WDR would have be NSW where, with a dispatch threshold of $10,000 MWh, our 1MW of DR would have generated approaching $95k in value and only be required for about 7 hours (over two years). So simplistically, $45k over 12 months. That’s certainly not the most lucrative way to use 1MW of flexible DER but it also doesn’t require a huge amount of effort to collect. Perhaps more of a cherry on top than the whole trifle.

For anyone interested in a closer look at the data we’ve published this to Tableau Public via this link. Please note this service is not maintained by us.

In some subsequent posts we’ll take a deeper dive into a couple of the topics raised in this post, including the qualification and baselining process – we’ll test some real-life energy loads from different types of business and see which make the cut – as well as the challenge of dynamic DER, such as battery storage. On the one hand, batteries should be able to access WDR and capture value (they’re fast!). On the other, the often dynamic nature of their behaviour, particularly if they’re ‘value-stacking’ creates challenges for the qualification step, but more on that later.

Data visualisation