Hide Accessibility options

+44 (0)131 625 1011

Search
Close this search box.
Podcast S21E05

Intelligent Demand

Big data connection technology concept.

In this episode, we speak to Centre for Net Zero, a non-profit, impact-driven energy research institute founded by renewable energy group Octopus Energy, we discuss the Centre’s work on intelligent demand, exploring recent experiences with GB’s Demand Flexibility Service, the growing role of automation, and CNZ’s work with open and synthetic smart meter data with Lucy Yu, CEO and Izzy Woolgar, Director of External Affairs.

Episode transcript

[00:00:00.140] - Jon Slowe

Welcome to Talking New Energy, a podcast from LCP Delta. I'm Jon Slowe.

 

[00:00:09.130] - Sandra Trittin

And I'm Sandra Trittin. And together we are exploring how the energy transition is unfolding across Europe through conversations with guests from the leading edge of the transition.

 

[00:00:20.040] - Jon Slowe

Hello, Sandra.

 

[00:00:21.180] - Sandra Trittin

Hey, Jon. How are you doing?

 

[00:00:23.270] - Jon Slowe

Good, thanks. Are you well?

 

[00:00:25.490] - Sandra Trittin

Also, well. Just out of Brussels today.

 

[00:00:28.830] - Jon Slowe

Well, Sandra, I was thinking back to demand-side flexibility, demand response, intelligent demand, whatever we call it, to a conversation I had many years ago, several years ago, with the then Chief Executive of SSE in Marchand. Super intelligent guy, fantastic leader in the energy industry. But he asked me, "Jon, how much flexibility will there really be from residential customers?" and it stuck on my head because someone Ian, a great leader in the energy sector, had that question. It was really bothering him, and I wonder if we'd really know the answer yet today. What do you think?

 

[00:01:12.120] - Sandra Trittin

I think there is no full, clear answer yet, but this has a specific reason because it's continuously growing. These distributed energy resources, either being heat pumps, being residential storage, being wall boxes, etc. This market is just growing and growing. And then for sure, you can do some market segmentation and you can have some thoughts around the concrete numbers. But how quick it will go and how fast we will be there, I think this is a bit of a different thing. And probably the second piece is also how much you get out of every asset can be quite different. For example, if you have a residential storage and you stop the storage from charging, you can say, okay, I would use this amount of capacity, but probably you do not want to fully stop it, but just reduce it, for example, and then it's a different game. So, you see there are different angles to it in answering that question, how much residential flexibility is out there in the market. We recently talked to Alex Schoch from Octopus. Today, I think we are really excited to have two people from the Octopus Centre for Net Zero.

 

[00:02:30.080] - Sandra Trittin

I would like to welcome our guests, Lucy Yu, who's the CEO, and Izzy Woolgar, who's the Director of External Affairs. A really warm welcome, and probably you can give us a short pitch on the Centre for Net Zero and also your thought on how to calculate this possible capacity for residential flexibility.

 

[00:02:56.430] - Lucy Yu

Yes, of course. Morning, Sandra. Morning, Jon. Thanks for having us on. So, Centre for Net Zero, why are we here? Well, technology is transforming the energy sector, yet modelling and policy continues to mostly be based on data from a past era. So, Centre for Net Zero, we design and run research and field trials around the world with the global Octopus Energy customer base to collect and then democratise data about the future, so what consumer behaviours will and can look like. And in doing so, we're really aiming to accelerate this journey to a fully sustainable global energy system. We're about 20 people. We come from a mix of different backgrounds. So, everything from data scientists, behavioural economists, strategists, and policy experts. And broadly, we do three things. So, we do the field trials that I talked about to collect, to gather that data. We develop and build and share really cutting-edge technologies, which I'll talk about a little bit more later. And they include grid models and some synthetic data work. And then we use the data and the insights that we gather to develop policy tools proposals that we think are easy to implement and can actually deliver a lot of the impact that we're aiming for at scale. So ultimately, we're really hoping to change the energy system for the better.

 

[00:04:27.290] - Jon Slowe

So the 20 of you must be pretty busy then. Yes. How I'm interested in the global nature of the work, Lucy, you talked about maybe Izzy or Lucy, you'd like to just tell us a bit about how Octopus is globalising, you're active in more and more markets, and how your work reflects that increasing internationalisation.

 

[00:04:54.300] - Izzy Woolgar

Yeah, absolutely. Thanks very much for having us on. We should say upfront at the beginning of the conversation, you mentioned talking to other members of the Octopus Energy Group, which is fantastic. We're a bit of an atypical organisation within OEG. As a non-profit research unit, we pursue change for the whole energy landscape, not just for one utility provider. We lead our own research agenda, and we operate autonomously. This independence was agreed right at the very beginning when Lucy was chatting with Greg Jackson, the CEO of the whole group, and they were talking about how the centre was going to work. Having that independence from day one has been really important for our credibility. The fantastic thing about being part of the Octopus Energy Group is, as you say, they are internationalising. Every week, we seem to launch in a new market. I think on last count, Octopus is operating in 18 countries. It's delivering energy services to 7.7 million households, and that number continues to increase. So that means that Centre for Net Zero has a real unique insight into the energy behaviours of people and businesses all around the world, and that's incredibly valuable for research and modelling purposes.

 

[00:06:19.730] - Jon Slowe

Thanks, Izzy. Lucy, in terms of the question that I, Sandra, started talking about at the beginning, and I remember the word Alex used, intelligence demand. Is that a word you use or how do you think about demand-side flexibility? Or what variation of the terminology sticks with you?

 

[00:06:42.510] - Lucy Yu

We do use the term intelligent demand here. And I think the key thing which we see is distinct from maybe traditional demand-side response, which was very static in its nature, really. So that would be aimed at quite crudely, maybe moving a large chunk of demand from one time period, for instance, to a different time period in quite a predictable way. But what we're seeing now, obviously, as we add more cheap but variable renewable energy to the power system, we need to be able to move demand much more responsibly, much more in real-time. That could be real-time price or carbon signals, for instance, perhaps applied in quite a granular fashion on the on the energy system, and to be able to respond very quickly, and perhaps in different ways to that. That's what we mean when we talk about intelligent demand. We're not really talking about moving demand in crisis, a fairly clunky and static fashion, but actually doing so very, very responsibly in real-time in response to external signals that could be changing hourly or minutely or every second even. And that could be changes into things like weather, which is obviously really quite unpredictable and very, very variable.

 

[00:08:06.040] - Jon Slowe

Well, should we pick up one of those examples? So, in Britain, we had an initiative called the Demand Flexibility Service, which uses old-fashioned terminology, not intelligent demand. But can you tell our listeners in a nutshell what that Demand Flexibility Service did and some of your key takeaways from it?

 

[00:08:31.800] - Izzy Woolgar

Yeah, sure. So, last winter, obviously, the energy crisis was taking place. So, amid heightened concerns about the security of supply in Great Britain and following a successful innovation trial run by Octopus and National Grid ESO. ESO launched this first-of-kind service called the Demand Flexibility Service. So, between November 2022 and March 2023, households were asked to turn down their energy consumption during one-to-two-hour periods in exchange for payment. Households were typically paid a fixed incentive which varied between events. From £2 per kilowatt hour to £4 per kilowatt hour. That was called the guaranteed acceptance price that was underpinned by government. Now, we saw 31 utility providers participating. Octopus Energy was the largest in terms of the total number of customers involved. There were 700,000 Octopus customers out of just over 1 million households that took part. They also provided the most amount of demand reduction. Rather than calling this the demand flexibility service, Octopus called it to their customers, saving sessions. So, Centre for Net Zero undertook large-scale statistical analysis of the 13 saving sessions that Octopus ran. And we also ran control trials whilst the DFS was running to better understand the parameters of consumer flexibility.

 

[00:10:08.280] - Izzy Woolgar

So essentially, our research showed a handful of really interesting things. First off, an important to say, households provided meaningful demand response, so system-level impacts. We saw a 40% reduction from those who opted into an event. Now, those effects are significant. They're comparable to a small power plant months’ worth of production. We did see that shorter notice period reduces opt-in rates and response as you might expect. So, customers...

 

[00:10:39.650] - Jon Slowe

How short, Izzy? What was typical and what was short?

 

[00:10:43.530] - Izzy Woolgar

So typical was day ahead. So, a customer might get a text or an email saying, hey, tomorrow we're running a saving session. It starts at 6:00 PM, can you please turn down your energy consumption during that time? When we saw intraday notification, ie. Within day or same day, even a couple of hours ahead of time, we saw typically around a quarter less demand reduction overall. However, there's obviously a trade-off between the amount of time that you give people and the financial incentive that you offer them. We did see a meaningful association between financial incentive and demand reduction. So, a £1.25 bonus saw a 4% lower demand.

 

[00:11:30.250] - Jon Slowe

Compared to?

 

[00:11:31.370] - Izzy Woolgar

On top of the additional £3 per kilowatt hour. Exactly.

 

[00:11:35.650] - Jon Slowe

Yeah.

 

[00:11:36.500] - Izzy Woolgar

Finally, we performed some really interesting analysis on the marginal value of public funds. For anyone not familiar with MVPF, that basically looks at the bang for buck of public policy. So, we found that for every pound spent on the DFS, there was a marginal value of public funds between £1.60 and £2.60. And that depended on the benefits quantified. It all came down to the value of lost load, ie. avoiding brownouts or blackouts, and that has the biggest impact. An MVPF of £2.60 is significant. It's much larger or higher than many other popular policy programmes that we see here in the UK, such as government spending on housing, job training, cash transfers, and health subsidies. That was really encouraging to see.

 

[00:12:33.300] - Sandra Trittin

What I found especially interesting and where I would be curious to learn a bit more about is this was a behavioural flexibility setup. You asked the client to take action. You were mentioning, for example, sending messages to them to reduce their demand. While we see in other continental European countries, it would hardly work, and people just prefer that it's fully automated. What do you think made people getting engaged so much into behavioural setup?

 

[00:13:10.860] - Izzy Woolgar

It's a great question and something we're very interested in here at Centre for Net Zero. So, we sent surveys to customers after they participated in the DFS and asked them exactly what they did during those turndown periods. And 75% of people came back and said they were manually switching off appliances. Now, that's great in the short-term, but as the energy system evolves to optimise demand closer to real-time, it's important to consider what the changing role of schemes like the Demand Flexibility Service might play, such as a important contingency resource targeted at times and locations where it is needed most. But in the not so far off term, we obviously need to transition to a system whereby consumers provide flexibility through smart automated low carbon technologies such as electric vehicles, solar pumps, solar PV, heat pumps, and batteries. This should happen seamlessly in the background without everyone having to think about it, delivering value both to households and to the grid at the same time.

 

[00:14:23.070] - Jon Slowe

Even so, I think the participation you got, you said 700,000 out of your 1 million customers took part. 70% participation, 40% reduction. I think it's hugely impressive. I know your customer base may not be typical of customers in Britain, but still, did that exceed what you thought you'd get, or was that in line with what you thought you'd get?

 

[00:14:55.420] - Izzy Woolgar

Obviously, we don't want to speak on behalf of the group, but I think the engagement was much higher than was expected. Obviously, it's a first-of-its-kind service, so you don't know exactly what's going to happen. There was this call to action.

 

[00:15:08.920] - Jon Slowe

Yeah.

 

[00:15:10.360] - Izzy Woolgar

Communication went out to customers. I think Octopus succeeded through really clear communication with consumers, explaining to people why this was taking place, how them participating would help protect the grids during times of strain, help wean us off fossil fuels. Simple messaging that really resonates with consumers. And there were also smart things that Octopus did. They ran gamification or they implemented gamification elements to the saving session. So, you got a streak if you participated in a certain number of sessions at the same time. They also offered things like philanthropic offerings. Consumers got an email at the end of their saving session saying, hey, you saved £11.97 in the last handful of saving sessions. Do you want to credit this yourself, or would you like to donate this to a charity of your choosing? And we will match fund that, ie. Double that amounts of money before then offering a charity. It's stuff like this that I think people actually really care about, and it helps really boost that overall engagement and participation.

 

[00:16:33.480] - Jon Slowe

Sandra, what's your... I think of the question that Ian Marchand asked me that I explained at the beginning and so, to me, the answer from this trial without automation is that there's a lot of potential. What's your thinking from what you've heard?

 

[00:16:50.690] - Sandra Trittin

I think there is a lot of potential. I just had discussions also with aggregators and demand-side providers from the US and I think it's a bit of a different setup, probably as well. There is no right or wrong. It's just different markets and different events require different actions. To give you one example that was just recently discussed was also that the dispatch, for example, to smart thermostats in the US is just a general signal that is sent to all the connected devices. So, you are one step further than the behavioural flexibility because you send an automated signal, but you just send a signal to everyone at the same time. And then out of statistics and calculations, you know roughly what will come back as some flexibility. Then there is a way where you can take it even further, where you really measure the behaviour of the device also within its context. You could even model, for example, a full home with a digital twin with a thermal model. And then you know exactly how this behaviour is going to change depending on the weather, on the client coming home on many different aspects. And once you are on that level, I think it's more increasing the flexibility that you can provide then afterwards.

 

[00:18:25.170] - Sandra Trittin

But there are certain occasions where it's just fine to send out a message and ask people to react, or where you just reduce the temperature, as I mentioned, from the smart thermostat by two degrees. But there are other markets or other market designs where this fully automated, almost like a closed-loop control setup is needed to maximise the flexibility and to make the business model also worth it.

 

[00:18:53.790] - Jon Slowe

Yeah, I think that I imagine the long-term future is more and more automation, and question as to how much behavioural is led on top of that automation. Lucy? Izzy? I don't know if you've got thoughts on that future balance between automation and behavioural. Will both co-regresses together? Will we move to having mainly automation?

 

[00:19:16.330] - Lucy Yu

I think you're right, Jon, that we will see a shift over time, in terms of the proportion which comes through automation, the proportion which comes through behavioural. I suspect there will still be a role for behavioural in the future even if it is a smaller role. I think in a way, the next 5-10 years are really about figuring out which elements of that behavioural flexibility we can convert into automated flexibility. I have spent a lot of time working on automation in different fields. I spent five years working in the automated vehicle industry. You would hear very often people saying, well, the technology is there. It's other things that we're trying to figure out. I think in the same way we have technology to automate distributed energy resources in the energy system, but we really need to do that in an intelligent way. This is coming back to this concept of intelligent demand. I think really this is about, first of all, understanding the very human side of things. So, we're not just applying automation in a very broad-brush way, in a one a size fits all approach, but we're really understanding what good automation looks like for different customers, different types of properties, buildings, and in different contexts.

 

[00:20:39.920] - Lucy Yu

So different times of year, different weather conditions, all of those sorts of things. And I think, really, over the next several years, a lot of what we're learning about through these field trials, which involve behavioural flexibility, will start to help us understand with a much greater level of sophistication, how much of that can be taken, can be converted into automated flex, and under what circumstances. So, I think we will see that move. And I think that will benefit everybody. And as Izzy said, as we said earlier, not all of this behavioural flex is particularly scalable or sustainable. Certainly, I think an element of it will always be possible, but a lot of it actually we need to operate really seamlessly in the background.

 

[00:21:34.410] - Jon Slowe

But it does show, doesn't it, what you described as very much customer pull. You developed a really good proposition... Octopus and the wider industry developed good propositions, good communication with customers, but there was a lot of customer pull. What I still see, some of in the energy sector, is energy sector push to customers, or I want to take control of this and control this, and it's not customer led. So, one of my big takeaways is how successful the demand flexibility service was at creating customer pull. And going forward, being able to harness that and learn from that as we get this shift, as you described, to automation.

 

[00:22:15.920] - Sandra Trittin

Yeah, based on that, one more question. How much would you think was the context of the situation of having this energy crisis? What was the role of this personal situation that we were all in at that point in time? Did it play a role or not really? What did your clients say?

 

[00:22:40.580] - Izzy Woolgar

Yeah, it's a great question, Sandra. It's something certainly we were all thinking about and mindful of. I mean, last winter, everybody was talking about energy in a way that I have not experienced before. You know family and friends suddenly all wanted to talk about energy and energy use, which was fantastic to see. So, I think people were much more mindful of their energy use. They were looking for ways to bring down their energy bills where possible. I guess, what we need to do is compare what happened last winter to this winter. The Demand Flexibility Service is recurring, it's been underway for several months now. And we are performing the same bit of analysis that we did last winter, this winter. One of the things we want to look at is developing longitudinal data sets when it comes to flexibility behaviours. And by that, I mean linking up the behaviours of a household that participated in the DFS last winter to this winter and seeing how did their flexibility behaviours evolve? Did we see fatigue? Was there higher engagement that first winter when there was the energy crisis, there were energy bills to consider, etc? Or are we actually seeing sustained flexibility behaviours?

 

[00:23:56.920] - Izzy Woolgar

It's a big question that grid operators and policymakers want to better understand. So, we're excited to be able to share those findings in due course.

 

[00:24:04.380] - Jon Slowe

Well, it's a good point to remind listeners of the show notes. Each episode has got show notes, and there'll be links to some of the work that Izzy and Lucy have been talking about in the show notes for this episode.

 

[00:24:17.090] - Jon Slowe

I'd like to move on now to some of the work you're doing around data, both, I guess there's a huge amount of data now available in ways that it wasn't available many years ago from the prevalence and rollout of smart metres and the rise of things we've talked about already. You mentioned digital twins, synthetic data. So, can you tell us a bit about how you're using these data sets and what you're doing with them?

 

[00:24:42.580] - Lucy Yu

As you say, Jon, if you want to be able to If you're thinking about energy policy, energy systems, and what's happening on the grid, you need access to data to be able to look at that. And that includes data about what's happening on the demand side. And so, one of the richest data sets to understand demand is the smart metre data that you talked about, so that half-hour consumption data. Ideally, though, as well as that data set, you would also have access to complementary data sets that just enrich it and help you understand it better. So, that might be things like the particular low carbon technologies that a building or property has installed. It might things like the energy efficiency of a building, so perhaps it's EPC rating and those sorts of things. I should say as well, actually, understanding what low carbon technologies are installed at a building alongside its smart metre consumption profile can also help you get a better understanding of what the trends that are occurring behind-the-metre. So, these are really quite hard to observe at the at the moment. So, if you just have the metre data, you are seeing what is happening in terms of energy being consumed from the grid or perhaps being supplied back to the grid.

 

[00:26:09.560] - Lucy Yu

But if you have, say, a household set up in which there are lots of low carbon technologies doing things behind-the-metre. So, for instance, solar PV on a residential rooftop and somebody is consuming their own, so they're generating energy and they're consuming it, then nothing has been consumed from the grid in that particular for example. So, there's a demand behaviour occurring there, but it's not observed through the smart metre data. So, data is very valuable here, the smart metre data, and ideally also other data sets. And we do know, we're aware that there is a movement here in the UK, and I'm sure the same exists globally, of different bodies, industry bodies, others, really trying to open up access to smart metre data. So, thinking about models for governance and sharing of that data. The reason this is not immediately easy and tractable is because that data is classed as personal data. So, it's governed by the GDPR regulations here and equivalent regulations around the world. So, it's not as simple as simply making it available for researchers or innovators or academics to be able to use. Now, we have looked at this landscape, and actually we think that there is a route to really enabling very widespread and global access to this type of data, but that route would be to pursue synthetic data so that's what we're doing at the centre for Net Zero.

 

[00:27:56.940] - Jon Slowe

Can you just unpack synthetic data?

 

[00:27:59.100] - Lucy Yu

Of course, we've built a generative AI model, which we call Faraday, and we have trained that model on nationally representative data set. At the moment, we're focused on our GB customer base, but this, of course, is a technique that could ultimately be applied globally, so localised to different markets. So, we trained the model on this representative customer data set, and it now outputs synthetic data. So, given specific household archetypes, you can output synthetic half-hour consumption data for those household archetypes. So, effectively, we are not only giving you synthetic smart metre data for whatever population of households you might have specified, but you also have that other data that I talked about attached to this synthetic data. So, you can specify, for instance, different local carbon technology ownerships for those households, different types of properties. You could specify a mixture of apartments and say, larger detached houses, for instance. And you can specify energy efficiency ratings of those properties. You can have an input which relates to the fabric of the building. We're also now working on adding in some more sophisticated handling of weather conditions as well, because we know that's an important determinant of energy demand.

 

[00:29:32.040] - Lucy Yu

So, what you get here is a completely synthetic data set, which is augmented or enhanced by those other things I talked about. So, the household attributes, the property type, the energy efficiency, the low carbon technology, ownership. We have been doing a lot of work now, and we're excited to be talking a lot more about this in the coming months. We've got several things in the diary already, but doing work to really objectively look at and think about how we assess and independently evaluate that data to effectively to check it is good quality data, but also to have some metrics that we can use to assess the quality of that data. When you have something like a generative technique that I talked about, you need to be confident that your synthetic data is not just regurgitating actual real-world data that it has been trained on. So, there are checks and balances that you can do to improve your confidence that that is not occurring, that what you're generating is genuinely synthetic data. And we've also been doing some work to check that that synthetic data maintains the diversity of the real data set that it was trained on.

 

[00:30:55.260] - Lucy Yu

So, we find that for demand data, particularly residential demand data, there's such variation in household and household circumstances and technologies there and those sorts of things, that this is a very, very diverse data set. It's obviously important that if you're going to generate a synthetic data set that it maintains that diversity. It maintains those very humanlike behaviours at the heart of it. We've been doing a lot of work to check that that remains present as well.

 

[00:31:27.580] - Jon Slowe

It sounds like, as you say, a lot to do to both develop the model Faraday or to produce the data to check and validate that and check that it's representative. Can you give a listeners an idea of some of the use cases or what you'd like to do with it or what you think others would like to do with it once you've gone through that process?

 

[00:31:56.880] - Lucy Yu

Yeah, well, I will hand over to Izzy in a second to talk a little bit more about our work with the Linux Foundation. But broadly, I think we see this as having a huge variety of potential applications. So, lots of planning related applications so, for instance, the grid operators, thinking about future scenarios. So, as more people adopt low carbon technologies, particularly at this distribution, the lower voltage distribution, distribution network level what does that mean for planning for where grid reinforcement might be needed or where it might not be needed if we can use flexibility, perhaps as an alternative? So, really understanding how demand profiles might change on the grid. Thinking about where to prioritise investments is another potential application of this data and all sorts of other policy applications. But we are about to launch an initiative with the Linux Foundation around us. So, Izzy can maybe talk a bit more about our wider ambitions for it.

 

[00:33:11.130] - Izzy Woolgar

Yeah, so for listeners who are thinking that data set sounds really interesting. I'd like to get my hands on it. We have good news for you, which is that we have recently partnered with the Linux Foundation, which is a global open-source foundation focused on creating techy We're focused on creating tech ecosystems to support rapid decarbonisation. With them, we are launching a global open data community called OpenSynth. And, really, the core objective of that group is to democratise access to this synthetic smart metre data. The idea is we want to empower both holders of raw smart metre data around the world to be able to generate and share synthetic data and for community members and energy researchers to generate, improve, and share algorithms as well. As you said, Jon, there's a lot of work to do, but the more people that are participating and sharing data, and the more that we can improve the algorithms that we're using, the greater the access to this synthetic data will be, and the more... Greater the access to this synthetic data will be, and the more use cases we will see evolving as well over time.

 

[00:34:32.640] - Sandra Trittin

Yes, and this will be then available, I assume, also to the community in terms of the improvements on the algorithms.

 

[00:34:41.090] - Izzy Woolgar

Yes, exactly that, Sandra.

 

[00:34:43.329] - Sandra Trittin

Yeah, that's great.

 

[00:34:43.500] - Jon Slowe

Sandra, I can see this question of understanding demand and goes back again to that question at the beginning of the podcast, how much flexibility is there from demand? Having, I was going to say real-life or synthetic real-life data, I imagine, will be very powerful for the businesses, for companies, for planners, for a whole range of people looking to understand value from demand flexibility.

 

[00:35:14.250] - Sandra Trittin

Yeah, understanding the value, but also understanding that a connected device is not... Or the capacity of a connected device is not necessarily the capacity that you can use for flexibility. I think this is already huge step, that if you have a 5 kilowatt battery, you probably do not want to use all the time these 5 kilowatt. And you cannot because at one point in time it's empty or it's fully charged. And the same with the heat pump. I mean, you cannot stop a heat pump for three hours in a row with the minus two degrees outside because the inhabitants of the home might think a bit different around that fact. So, I I think that will be already a huge step forward also in terms of thinking from the perspective of the customer, in terms of how much capacity or flexibility will be there to control. I think it's in the end, the equation of the total number of assets within their own context and how much you can get out of each of these assets at one point in time. This will...

 

[00:36:27.780] - Jon Slowe

Yes, it's dynamic over time.

 

[00:36:29.570] - Sandra Trittin

Exactly. And how Izzy and Lucy were also explaining, this is continuously changing. Because you're coming home, you put on your lights, you might turn up the heating because it's colder today. It depends on so many different constraints that you need to manage them as well. And I think this will be then the future of the, let's say, management of this demand, of this intelligent piece of the demand management, that you need to have a model or a system which always adapts the algorithms continuously, based on these changing parameters on the model.

 

[00:37:10.780] - Jon Slowe

It sounds exciting. It sounds complex, but I think with the data tools that we have. Lucy, how hard does that work been to develop these synthetic data profiles? Have you really pushed the envelope, or has it actually been not too hard once you got stuck into it?

 

[00:37:29.210] - Lucy Yu

You might get a different answer if you asked my team, but we've been working on this in the background for probably about 18-months now. We began actually by exploring some slightly more popular generative techniques. And we landed on the one that we did because we found it better to maintain the diversity of the training data sets. As I mentioned, that's important because there is a lot of variation in the way that households consume data. But I think the nice thing about this is you do a lot of the difficult work up front, almost, and then you have a framework or an approach that we can then take, and we could, for instance, begin to train this on data from completely different countries. So, we could begin to localise this data set. And as Izzy talked about, the community that we hope to build with LF Energy, we will also then be able to get contributions from others globally and perhaps even improvements to the model that we're using as well. So, I think this has the potential to scale very quickly in a way that I'm trying to open up access to real smart metre data.

 

[00:38:54.720] - Lucy Yu

Just inherently, it's a very one to one mapping. I I think the other thing, we are seeing some smart metre data begin to be released in the UK. So UKPN and SSEN, two of the distribution network operators in the UK, have recently begun publishing some low voltage reader, smart metre readings. This is useful, and I'm sure there are things that we can all do with that. This is still aggregated to, I think, a minimum of 5-10 households. So, it is not that individual household level, and it doesn't have that metadata that I described. You still don't have that information about the things like the technologies that those households own. I almost describe this as a bit like the supermarket trolley problem. It's being able to see what's in people's trolleys but knowing nothing about the customers themselves. So, that probably has some usefulness. If you're running a supermarket, it gives you some information about maybe how to think about supply supply chain and operations and other things. But because you don't have all of the data points, you may still make some decisions that actually are perhaps not the best decisions that you could do if you just have that little bit.

 

[00:40:13.870] - Jon Slowe

That's a nice analogy. Well, best of luck in taking that work forward, sounds very exciting. Time's getting the better of us. So, let's bring up the Talking New Energy crystal ball, and I'm going to set the dial this week to 2030. Six years away, so far enough for a lot to happen, but close enough that I've got to ground it in realism in the next years. My question to both of you, looking back from 2030, how would you describe the progress we've made to locking the full potential for intelligent demand or how much progress will we have made. Lucy, do you want to go first, and then Izzy, maybe add any thoughts?

 

[00:40:55.340] - Lucy Yu

Thanks, Jon. Well, I think it's a relatively short timeframe. I I think we talked about how flex has real value in the energy system today. We talked about our analysis of the demand flexibility, and in particular, our findings about the marginal value of public funds. I think, as Izzy mentioned, the results that we saw there, those are really striking results in public policy terms, and that already implies that this is an option that we should be giving much more attention to. I think it is clear that flex would have even greater value, particularly in the GB energy market, if that market had a slightly different shape and if we made more use of granular pricing structures in the market. I think by 2030, we will recognise that actually having a more sophisticated market design can actually help us create more value from flex. And I think when that happens, what that means is we will also see just more innovations, more and more organisations coming into that value chain to capture a greater proportion of that value. So, really what I'm describing is I think there'll be a bigger pie overall in terms of the value from flex, and I think more of that pie will be captured. I think we will see that returned.

 

[00:42:16.760] - Lucy Yu

We'll see two returns on two levels. We'll see returns to households then in the form of lower bills, and we'll see returns to the people, the organisations in an industry who innovate the most successfully. We always see this when markets change, new value is created. Those who are strongest at innovation tend to be those who capture that value. And my hope is that we have a lot of field trials underway at the moment which are squarely looking at all of these topics around flex and intelligent demand. We would be extremely pleased if the results and the data and the insights from those field trials have gone in shape policy to ensure that those benefits that I'm talking about and those cost savings that I've talked about are really being spread across all households. So that's my 2030 crystal ball goes in.

 

[00:43:13.000] - Jon Slowe

Thanks so much, Lucy. You think?

 

[00:43:14.980] - Izzy Woolgar

Yeah, I mean, I hope we continue to make a lot of progress. Certainly, in the last few years, we've gone from in the UK, thousands of people participating in flex, and now we're seeing millions. But there's still an uphill mountain to climb. We've got about 6 gigawatts of flex in the system at the moment. By 2030, we need 20 to 30 gigawatts. By 2050, we need potentially over 100 gigawatts. I think a lot of that progress is going to come down to the innovations that Lucy described, both from market players, but also from the way that we design energy markets. But I think a lot of it will also come down to political will to communication. Sandra, you're sat in Brussels today, and we've just seen before Christmas a failure to introduce significant and ambitious policies across 20 plus different countries in the form of the Energy Performance of Buildings Directive. So, European legislators have reached an overwhelming deal on this. It's nowhere near as ambitious as everyone was really hoping it could be. Ultimately, that was diluted by conservative governments and groups in the European Parliament. Now, why did it fail? It failed because of misinformation and poor communication.

 

[00:44:41.960] - Izzy Woolgar

And, what I really hope to see in the next six months is to 2030 is far more clear, coherent, and really evidence-driven conversations about flexibility, making sure that it's incredibly clear the value that it can deliver to consumers and to the energy system and I think that's what will really turn the dial moving forward.

 

[00:45:03.590] - Jon Slowe

Lovely. Well, I share both of your desires for the next six years. I think if we can achieve that, then we'll be a lot further forward in unlocking the potential of flexibility and intelligent demand. Lucy, we better draw it to a close there, keeping on time. Thank you both very much.

 

[00:45:25.320] - Izzy Woolgar

Thanks very much.

 

[00:45:26.880] - Sandra Trittin

Thanks a lot.

 

[00:45:27.770] - Jon Slowe

Sandra, any reflections from your side?

 

[00:45:31.220] - Sandra Trittin

I still keep on being excited around that topic. The more we talk with different people around the topic of flexibility, the more we see the different perspectives and also how we can all grow together. I think this is one big learning for me, again, from this episode that we see it can only work together in collaboration. Now with that open data model, I think I think that's also one example. Again, we need to work together to make that happen, and we will not make it one-by-one. This is one key element. The second key element where we are coming always down to, Jon, in many different topics in our podcast is communication and education. I think probably we should start a communication agency or something. It's not only on the topic of flexibility, but everything that's around these renewables. We have seen that also in some European countries when the legislation was supposed to change on how many facts or non-facts were raised in the communication also towards the end-client, how much you can reach or how much you can also miss with the wrong or right communication, the influence of that. So, these are two pieces that really stood out to me today again. How about you?

 

[00:47:06.360] - Jon Slowe

I think I would take that communication and education piece. And my takeaway, two takeaways, that the customer interest can be harnessed. I think that's clear from the GB example and many other international cases. And the market design is so critical. Using the evidence base, the communication to the market design, I don't doubt if we get the market design right, that there'll be any shortage of innovators in the market able to excite customers, unlock value, create value for themselves as innovators, share that value back with customers as Lucy described. But we got to get that market design right, and that requires a communication and education and collaboration that you talked about.

 

[00:47:59.750] - Sandra Trittin

So, let's get going, right? Let's make it happen.

 

[00:48:03.380] - Jon Slowe

Yeah. Keep pushing. Okay. Thanks to everyone for listening. We hope you enjoyed the episode, and it's given you a new perspective, the new things to think about, about intelligent demand, flexibility, and open data. Thanks very much, and goodbye.

 

[00:48:21.250] - Sandra Trittin

Thank you. Bye. Thanks for tuning in. We are excited to bring you captivating conversations from the leading edge of Europe's energy transition. If you got suggestions for topics or guests for future episodes, please let us know.

 

[00:48:34.650] - Jon Slowe

If you're enjoying the podcast, then please do rate it and share it with colleagues. For show notes, transcripts, and more, please visit lcpdelta.com.

Add yourself to our mailing list

Add yourself to our mailing list