The Digital Backbone of Energy Storage Optimization with Roger Hollies (Arenko)

I'm Ed Porter, VP of Insights at Moto Energy, and you're listening to Transmission, the podcast all about the world of battery storage. And often, batteries are still talked about as simple things.

Charge them up and discharge them later. But as today's guest Roger Hollies explains, that's not how they actually work on a modern grid. Grid scale batteries respond on a microsecond scale. They don't just shift energy, they help stabilize frequency, smooth voltage and correct tiny imbalances before they turn into bigger problems, which means they're not passive.

They're active control systems. Roger is the CTO of Arenco and the team delivering Nimbus, Arenco software that controls large battery portfolios in real time. Once you start operating batteries at scale hundreds of megawatts across multiple sites, the challenge stops being the hardware, it becomes coordination via software and in this world AI enablement is moving fast.

But it's also about making sure that market rules can keep up with machines that move far faster than humans because poor enforcement raises costs for consumers, good actors get undercut and ultimately confidence in markets erodes.

The UK has world leading flexibility markets but enforcement lag behind design and we could be moving so much faster.

Roger argues that poor delivery has sometimes gone unpunished and bad behavior distorts outcomes. So how can market rules keep up in this new world and how can evolving software help control these assets to best support the grid? And how do we instill trust in automated AI driven software that lives inside a black box. Welcome back to transmission. Let's jump in.

Hello Roger and welcome to transmission.

Hello mister Porter. Thank you for having me.

Thank you for the very formal introduction. And as ever let's start off with who are you and who is the company that you work for?

Great. Yeah. My name is Roger Hollies. I'm the CTO at Arenco.

Arenco are a technology company operating in an energy space. We've been in the space now for ten years.

Our Nimbus platform is a software technology which helps large energy companies manage global portfolios of batteries and renewables.

Okay. And maybe just give us a bit of context on like how many regions you're in and sort of the megawatts or gigawatts under management. Yeah.

So we've got about one point two gigawatts of batteries operational and that's batteries including batteries co located with solar, batteries co located with wind and that's predominantly in the UK.

But also we have a footprint in Texas and Germany and Ireland as well. Okay. So yeah, a little bit spread out.

And does that have a bit of a range?

Right? Do you have some really big projects and some much smaller ones?

Yeah. So starting from five megawatts up to anything up to a hundred megawatts and larger coming online soon. So it's across the range of grid scale energy storage and renewables.

And the role of Orenco in the market, I know it's kind of changed over the years. So do you wanna give a bit of context for sort of what you've been looking at, but also where you're the, you know, where you are headed?

Yeah. So you as I said, we've been in the industry now for ten years. We started out when there were no batteries plugged in to the to the network, so that was back in two thousand fourteen, two thousand fifteen. And we've grew through developing the sites, putting in some of the first batteries in the UK. We had a one megawatt trial battery back in twenty fifteen-sixteen that we were trading with and providing frequency response services.

Then on the back of that we raised money to build a forty one megawatt system which at the time was the biggest battery in the UK.

Alongside that we developed the technology to control it because nothing we were seeing in the in the market was able to do the complex time shifting of energy from from low to high and then the host of power quality services that you can provide with batteries that we could see would be a major component of it. So we developed our own technology to do that. We started offering, started operating on our own asset base and then realized that it was really that software technology that was the USP of the company. So moved on to providing that technology to others in the form of route to market services, but we were really focused on from a very early point of offering software.

The tools in the Goldrush kind of concept, right? So, you're building the tools that companies need to operate this this wonderfully complex asset class into an increasingly complicated and changing kind of market construct. So, the energy industry as you know, it's just change, right? That's the one constant change in complexity and maybe two.

So we've developed tools to abstract and simplify and particularly for large any energy companies to manage those big portfolios into that increasing complexity.

Okay. So so it's both you started off very storage focused Yeah. And you're now looking much broader. Is that sort of fair to say? So you're more looking at either the the portfolio of assets that a large energy company might have.

Yeah. Exactly that. And the way the reason for that is mainly customer led. So we've done a lot of work on hybrids, co located batteries, located solar, co located with wind, so a shared good connection and sometimes some shared commercial constructs. And because we started off with batteries which are by far the most complex asset class in terms of particularly for their scale, you know, are lots of little power stations have to be orchestrated to one power point, they're measured at twenty hertz, you know, some cases more in terms of their power response. So we've start off that really complicated end and so we're actually moving that technology to deal with some of the issues with renewables is actually quite an easy step.

And it's been a pull from our customers who have used the Nimbus platform.

It's been a pull for them to being actually this could be really useful on our renewables fleet.

Maybe maybe two questions on that. So when you see something is measured at twenty hertz, what does that mean? But maybe just to give people an idea of like the number of decisions being made in every say second by Nimbus. Well, like, does that look like? Is it sort of one or is it a hundred or is it ten thousand?

Yeah. So that's a great question. So measuring twenty hertz, so twenty times a second, so you're measuring one data point every fifty milliseconds. So twenty times a second and you're collecting that data up but also responding to frequency change with power. So frequency response is managing the frequency on the grid. A lot of your previous guests have have talked about it. But it's one of the key kind of demonstrators of batteries capability because they can do it incredibly well, incredibly quickly, incredibly accurately.

In terms of decision making, so we collect so our our software does a number of different things but the core things are control and dispatch. So we have hardware that sits on the site, integrates with the asset, controls and dispatches it, we collect a huge amount of data from those assets and then we do the optimization into the markets as well. So there's an asset management control and dispatch asset management and optimization.

We collect at the moment with that one point two gigawatts of assets we're collecting seven billion data points a day.

So it is a vast amount of data that we're collecting and the key thing about our software is to take that data and make it usable either through automation or through insights into into into what how you can improve the operation of that asset. That's on the asset management side. So it is yeah. It's a huge amount of data, a huge amount of compute power going into that to try and to try and manage these assets into into the future state.

And and just to kind of recap on that, so you've got sort of the very basic decisions of charge or discharge. So, you know, megawatts of energy in or out. Then you've got the reporting element, which is how much of this gets fed up to the the asset owner and, like, what do they see on their screen.

Yeah.

You then also mentioned very very sort of quickly something around sort of like broader power quality. Yeah. And and I think we're going a little bit into detail here, but but there are more things that a battery can do than just sort of energy and energy out. Yeah.

And you mentioned one of them in frequency response, which is actually energy and energy out. Yeah. But that is kind of an example of something else you can do rather than just sort of the blunt shifting of energy. Yeah.

So are you thinking about other services that these might be able to do? I'm thinking of things like reactive power, voltage control. How close are you to those those services?

Yeah. So the way I the way I describe what a battery does is you've got the basic time shifting of energy which is from the peak and that could be if you align it with solar, you're wanting to take power from the middle of the day and then shift it into the evening and evening peaks. Then I like to I describe it as the suspension of the grid. So electricity that comes out our plugs and and powers industry has to be within certain parameters for its frequency, how quickly the the waveform travels, its voltage, its inertia characteristics.

So there's this whole host of kind of power quality that you have to adhere to. And batteries are very good because batteries basically define the waveform, they define the electricity that comes out them really accurately. They can manage that power quality and deliver really smooth power or jerky power whatever is needed in order to maintain power quality at the meter. So, it's a really important part of what batteries do.

And from day understood what batteries could do in terms of this managing the waveform and the power quality. So all of our systems are built again trying to abstract away from things like a frequency response concept. What you've got is a signal coming from the measurement, which is the frequency of the grid or the rate of change of frequency and then a response in power. But that those measurements could be voltage measurement and reactive power, both of which the react power is another component you can control with batteries.

So we tried to abstract away from specific markets and capabilities and just try to maximize and this is all come comes back to the change point, the one constant is changing in this industry. Very early on we knew we didn't know what we were going to be using the batteries for in two or three years' time, so we just built as flexible battery and then a flexible as control system as possible. And yeah, you've got markets now emerging around voltage thermal constraint management, the inertia, big push at the moment for inertia, black start, all of those things we have tried to build the constructs within the platform in order to take advantage.

Yeah. And then that becomes very useful as well if you are then looking to go into different markets, which might have because at the moment our trading and optimization strategy is all based around the UK. But power is power. Globally, these assets assets are the same assets that are going in globally, but how those markets are managed is very different.

So to have those constructs abstracted, it means that we can then lift and shift our technology into these different regions.

Yeah. All of the rules are different by country, but the fundamental of like batteries, solar demand, it's all pretty similar by region with some sort of geographic differences. I think maybe one exact one sort of way to think about the the energy balancing, the frequency balancing, the voltage, the reactive power, those those two are linked inertia. Imagine someone sort of spinning plates in all of these markets.

Yeah. You've got something like twelve to fifteen plates that you need to spin to manage the grid successfully. Yeah. And batteries can do a huge number of those.

And what you're saying is that you've kind of got the flexibility to bring batteries into each of those markets.

Yeah. Exactly right. And, yeah, offer them the services, generate revenue for the owners But also contribute to the problem that I think most of us are in this industry to solve, which is the move to a sustainable energy system. The increasing the ability for grids to take solar and renewables. When we started ten years ago, when I studied fifteen years ago, there was like kind of an understanding that you couldn't have more than forty percent penetration of renewables into grids. Now we're seeing days where we're going to eighty percent to ninety percent of renewable of energy being provided from renewables. So those again, all those rules are getting thrown out the window.

So really exciting time to be in.

And to and to tie some of that back in as well. So if you look at one of those like really sunny days in the middle of summer, often it's like a gas unit that's running and it's running to provide voltage control. So like one of those plates that we were talking about spinning is the only thing that's kinda keeping some gas running. And actually, you've got the right control system and the right asset, you might be able to provide that service with a with a battery. Yeah. And so the hope then is that you can get from say eighty to ninety percent renewable penetration or ninety to a hundred percent.

Yeah. So I mean, again, and I'm not gonna try and not bash gas too much but at the end of the day we're trying to keep the fossils fuels in the in the ground and not burning into the air. But over for the past eight years since you introduced markets for frequency response, EFR, enhanced frequency response was brought in I think eight years ago two thousand eighteen?

Yes. Yeah. Depends what like yeah. Yeah. So REGS REGS was twenty sixteen. Marketing went live in twenty eighteen. Yeah.

So since then you you went from having zero batteries on the grid providing frequency response to now where ninety five percent ninety ninety five percent of it is being provided by batteries and this is the stay of the rate of change that's happening in our industry. You've got a brand new asset class coming in and dominating a critical part of supporting grid infrastructure. And it's done that over a really short period of time displacing gas and that we see is happening across reserve markets.

Basically batteries capability sets will stretch out and dislodge the incumbents mostly which are fossil fuel Yeah.

And and using clean energy, the clean generated energy to do it. And again, this is why it's such an exciting industry to be in.

And we've said this before but if you're a system operator and you're you're hearing this like, the open markets is kind of the key thing. So we would say, don't don't take our word for it. The batteries are the best thing for any of these things. Make an open market, allow people to compete, and then see what happens. Yeah. Our money is that batteries will do a very, very good job. But equally, if the if the particular problem that you've got is a problem that's solved by gas then Yeah.

Fine. And the defining the market is an is an interesting point. It's about defining the problem. I think this is what Niso have have got very good at in the UK is getting better at just defining the problem rather than saying, right, we need this market. They're like, actually, we've got this problem, what can solve it?

Yeah. And NISO are the national energy system operator. Yes. So they're in control of spinning those plates, which is why they should be defining the problem.

Professional plate spinners. Professional plate spinners. Yeah. I I think I don't know whether they'd be happy or unhappy with that.

We'll we'll find out. Okay. So we've just gone into a little bit of of the depth of the market, but I think really important, right, because it explains why this sector is really important and why each of the individual parts kind of help to contribute to essentially making the wider system work.

Quick break. If you listen to this show, then you probably work in energy. And at Modo Energy, we're not just talking about the energy transition. We help our users to actually make it happen.

All energy storage, solar, and wind assets on the global balance sheet need to be valued and benchmarked, and that's where Moto Energy comes in. Our benchmarks and forecasts are transparent, bankable, and trusted by the world's leading banks, asset managers, utilities, and developers. So if you want to learn more, go to modo energy dot com. And if you want more content like this, sign up to our weekly dispatch newsletter.

Enjoy the conversation.

So what what's really interesting for me is that Arenco have gone from really being in the detail of, like, how exactly these systems work and say spinning twelve or sixteen plates to also then looking across sort of full portfolios, looking across wind and solar. And to me, that sort of portfolio management feels more like Kraken, which is obviously one of the tools that has done incredibly well in in GB over the previous years and is something, I think you I think I'm right in saying is a double unicorn in terms of a a software company. So a huge success story.

Would would you see parallels between Arenco and Kraken?

Yeah. I mean, I like any kind of parallel that puts me against a ten billion pound company valuation. Look, Kraken, I I love Octopus, Kraken, that whole construct. They've done amazing things with the industry in terms of making it attractive and kind of sexying it up.

Our focus is on grid scale technologies, so wind, solar, batteries and controlling those. Part of Kraken's remit is doing that but they also do the behind the meter and the CMS system which is their major drive. So whilst they're a wonderful kind of David and Goliath style competitor for us and we go up head to head with them in a lot of the RFPs request for proposals. So there are lot of bids that come up from the energy companies.

Yeah, definitely there's a lot of overlap not least there of high functioning technology company operating in the energy space, right? That's what we do. You can't have monopolies in industries, so we kinda welcome the competition. I've sat down numerous times with Charlotte over over tea and drinks to try and figure out, you know, Charlotte Johnson from from Kraken.

Trying to figure out if we are competitors or not or if there's spaces around and the reality is

It's quite nice to realize it and still be friends that we're just straight competitors in in some of the work we do. But that's good. I mean, that's really good for us to have that kind of level of competition in the industry.

And our tech's better.

Yeah. There we go.

If I'm allowed to say that. I think but you know, that's feedback that we're getting. I'm sure, yeah, be it you know, there's there's areas that they and we know that they they build really really good tech but we know our tech is really really really good.

And if you want to delve back into the archives, think Charlotte might have been episode five on on transmission and this is probably episode two hundred and twenty or similar. Yeah. I don't know exactly, but maybe a space to kind of pick up pick up that challenge on who's got the better tech. Yeah. Definitely. In the future?

All day long. Happy for that.

Okay. So so maybe let me ask that in another way. So in five years time, what's the perfect position for a Renko in the market?

Yeah. So hopefully we continue the the mass build out of renewables and the continuation of evolving evolving grids, electrification of cars, transport, driving more and more requirement for flexibility. And I think what we would like to be is a recognized kind of global tool and a kind of digital backbone for some of the big energy companies who have evolved and have managed that transition to managing these huge complex portfolios. And our tech is at the heart of improving their efficiency, allowing them obviously to make more money, but also allowing flexibility systems to be more effective, to be managed in the most efficient way so that they can deliver their services, improve how systems operate and key to it all is managing that cost to think consumer, right?

Yeah. So we've got to be delivering the just transition. The energy transition has got to work, it's got to be secure, it's got to be clean and it's got to be affordable. So, I think that's ultimately we want our tech to be in the backbone, in the detail, we love complicated problems and solving those problems with with mature digitalization infrastructure.

Okay. I love that. Let's let's move into the market space. So when we talk about grid scale energy storage and automation, what would you say the sort of number one thing is that people misunderstand about how that works?

Maybe about the complexities of automating batteries is a wonderfully and a step change in complexity. And I think the reason behind that is the batteries at the very basic level can charge and discharge which is unique. It's not a demand, it's not a generation, it's a bit of both.

But there's also concepts around it in that a, the duration is commonly misunderstood. So a, there's been a lot of talk about the long duration assets with the recent release of the long duration energy storage plan consultation and markets.

But a two hour battery is a four hour battery is a eight hour battery just sort of lower power. And I think this dynamic shape of a battery and to give a really firm example, a hundred megawatt two hour battery is is a battery that can charge or discharge for two hours at full power.

A hundred megawatts if you charge or discharge at twenty megawatts could be going for ten hours in theory charge or discharge given the efficiencies. But that again, that flexible construct of the batteries trying to automate into markets is just one of the complications of how complicated it is to abstract it up and particularly into traditional trading desks. So one of our, one of the key pathways that we see our technology providing and it's been proven out with contracts with the likes of Orsted and RWE is most of these big energy companies are bringing one or two batteries into their portfolio and they trade or operate gigawatts and gigawatts and they're bringing one fifty megawatt asset in, battery asset and it's completely different to anything they've operated before.

So they need some kind of like initial stage of tech to help with that first operational cycle. And then once they start getting to scale, they can use our complete end to end trading automation or they can start peeling back and say, well, actually we've got our own forecasts. We've got our own market optimization engine that we developed for batteries and we want to start using that.

So we allow for that journey and again, one of the characteristics of our platform we've developed it in a very modular kind of way is that people can kind of chop and change with the capability sets that they use. It's like a Swiss army knife, nobody ever uses the funny thing with a hole in which nobody knows what it does, but it's still there.

Our platform doesn't have anything that nobody uses because I don't know what it's Funny thing with a hole in it.

Yeah, definitely.

Now I want to know what the funny thing with a hole in actually is for sure.

I some of my engineers would definitely be able to tell you what that is.

But the idea is that you have a platform which has the ability and is built in a secure redundant enterprise grade way that can do all of these things. Actually as a global energy company, have IP in forecasting, I have IP in optimization. So, want to use that IP to to manage my fleet. But it's having a flexible kind of technology partner that allows you to kind of move into that space. So, possible. We're not trying to build out tech and all you can use is a Renko.

One of the core constructs of our platform is that it plays well with others.

Yeah. We store all our data in modern data lake infrastructure with modern API connectivity.

So other digital participants be that our customers or people who who our customer contracted with, they can use Yeah.

The data.

I think that was definitely something I wanted to ask about which is when you go to a battery site today, is it sort of the Nimbus system and then the client systems? It's like a one to one relationship, or is it more like there's the client over here, the Nimbus system sits in the middle, and then to take just an example, there are people like Twice or Acure who do, like, battery state of health work. There are others as well. Yeah. But effectively, they would also slot into Nimbus as well. Is it more of like a sort of hub and spoke type thing, or is it what what what does the relationship look like?

No. I I think you've kinda nailed it on the head is that what we're trying to do is develop a system that integrates with the site, pulls all of the data alongside any other providers who are on the site but more so the likes of Twice and Acure and other providers who provide similar battery analytics. They really don't want to be installing kit on all of these sites and you know, and that can be the same of any other provider into this market. So we provide that plug for people to basically get customers to get their data to people who want to use it off the site. We do think having our hardware on the site at the moment is really important because we have our own metering. So, we get that power quality and we can understand that power quality and that source of truth of what's happening at the meter, which is really important.

Okay. And now I think probably to the question that everyone kind of wants to ask and we get asked a lot which is to what extent is optimization a manual job and to what extent can asset control rely upon AI?

Yeah. So this is a a really interesting area. We've been using AI in various forms for forecasting and optimization for like six or seven years and it is imperative that as an as an industry we use it, it's a productivity tool, it's tools, there are problem sets that it really suits.

There is a an interesting kind of mad rush right at the moment, the AI bubble or whether it is a bubble or not, but it's certainly certainly very, very exciting and very unchanging at a rapid pace. But it's all about using the correct AI that most of the hype is around kind of large language models, right?

Things like chat GPT.

Now, and there are applications in the energy industry where that works. But on the whole, there's there's appropriate tooling for for the job. So, I like in the the metaphor I use is a the kind of band metaphor. So, you've got a band but you don't wanna have like a top level jazz musician playing in a wedding band because they will just shred all over everything and make everything sound like just really nasty and they're amazing and they're incredible, great at what they do, but they're not appropriate for the use.

So I had a great quote the other day, kind of AI thought leader was asked by a corporate level, should we get an AI, a head of AI in? And he said, well, do you have a head of spreadsheets or do you have a head of PowerPoint? No, you just train people up in the appropriate tooling and that really is what AI is, it's a tooling that can increase productivity or on the front line of development like software development, So, there's tooling to support with software but you can build AI models into your kind of tech stack and use them. And it's all about training people up to use the appropriate AI for their role.

So, that's how you think about it broadly. Yeah. If we make it sort of make it real in terms of saying this asset a is going to be dispatching for this evening, how often is that sort of a person doing that and how often is that AI saying, I think it should be this and then someone kind of approves it?

No. So so with the way we operate and without giving too much away, we have multiple kind of different layers of use of AI we have in the forecasting. So obviously with batteries you need to understand with operating batteries into any markets, you need to understand of the battery and you need to understand the future state of the market, both of which are unknowns. So with the future state of the market, you have price predictions that are using things like neural nets.

With battery future state, you've got again self learning models which are looking at past of how the batteries behaved and then looking at the future current state any trades that place so you understand that.

Then when you have those two pieces of information and lots of other broader market kind of information, you can then optimize and say right, to make the most money over the next twenty four, thirty six hour period and that's the time period we're focused on really the next day and intraday.

You're saying right, what's my pathway of max revenue?

And with our system every minute the optimization engine runs, we're working on increasing that but actually with a resolution of half hour trading windows it's probably enough. Every minute we're taking the latest information and running an optimization engine which is basically saying right this is my new plan, this is what I want to do, these are the mark the prices I want to place in. At the day ahead stage we use that model to give a opportunity cost, So, we're bidding in all to all the day ahead markets based on an opportunity cost. But again, we're using kind of automation, not AI but automation tooling to like build, build and build bids and basically try and maximize revenue at the day ahead stage based on the opportunity cost. So there's kind of like multiple layers of use of automation alongside AI and AI is just part of your kind of automation tooling, it's the way we look at it.

Okay. And then overseeing this there is usually someone looking at the assets and saying, yeah, okay, I like that as a strategy, let's let's go ahead and do it.

Precisely at the at the strategy level is where you need your optimizers and your traders. Yeah. And in some instances, so one of the things we've worked really hard on particularly with AI, one of the problems with optimization engines is the black box nature of them and perfect hindsight is a nightmare in in trading because you look back and say, why did he or the optimization engine, why did they these trades when they could have made these ones and it would have made loads of money, made more money.

It's what we've worked really really hard on and we have released some capability on this and this is gonna be probably by the time this episode is released, we'll have it out in the in the main platform for our users, it's currently in beta form being used internally, is a visualization of the decision making. So at any of any point of those one minute windows historically, you can click on click in and see exactly what the optimization engine was seeing and what its plan for the next thirty six hours was. So, that is constraints on the battery, constraints by services that you've committed to provide, so state of charge limits that some services apply, any trades that have been placed, what the price forecasting it's seeing at that moment in time is and any other kind of external constraints like like wind constraints if it's co located with wind or solar, there might be import or export constraints due to something externally.

So, for every second you can see basically why why it was doing what it was doing and why it placed certain trades here and there. And therefore break open this black box concept. And if you have all that information, you talked a little bit I think about trust in automation. Trust comes when you understand what the automation is doing.

Yeah. And this visualization of this automation engine is is really cool because you can then see now you know why it did what it what it did at any moment in time. Yeah. Then you can disagree with its forecast or disagree with its logic.

And both those things are things you can improve if you've got data on.

And so if you're also an asset owner, you might see two prices from yesterday. One might be a hundred pounds and one might be a thousand pounds. And it looks like your asset discharged into a hundred pounds. And you go, what what you doing?

Like, why does he not pick up a thousand pounds? And you can then say, well, actually, let's look at the decisions we had at two o'clock yesterday afternoon. Exactly. They said the peak price was gonna be a hundred pounds.

So we went for a hundred pounds. After that period, a big nuclear asset tripped. And all of a sudden, there was chaos on the grid, and it went to a thousand pounds. But we only had one megawatt to sort of sorry.

One megawatt hour of energy to put into the system. We put that in at a hundred pounds, and once we've made that decision, we were already locked in. So we we couldn't do something else at that point. And so you're saying, look, as long as we can provide trust and transparency around decisions we make, then we hope that asset owners can understand how we got to that point.

Exactly. Exactly.

Okay. One last question on market. So what is the sort of number one issue that's sort of working or not working in GB today?

I listened to your episode with Lisa McKay and from Fidro. And her contrarian viewers don't bash Ni So.

And in which I completely agree with. However, it doesn't mean we can't ask them to be better. I think we've had a historical problem with policing of markets, think.

In that they've created some of the most sophisticated frequency response market, reserve markets and open development with industry of markets, you know, forced upon us in some regards because we're an island state, so we need to have a really solid grid. But in terms of making sure people deliver these services correctly, I think there could be a huge amount, we could have done a lot better in terms of policing them. And there's a post I made after back in twenty twenty one, eight weeks after dynamic containment was released, I'd said, right, eight weeks, how are we doing? And I kind of plotted, it was really proud of my of our performance. It was like eight weeks of and this is one of the markets that's measured and penalized that twenty data points a second. And I'd had just demonstrated that we were nearly faultless for that entire time.

It's not beautifully perfect. And if you do, if you do, you know, don't hit the response correctly, you're measured and you're penalized using something called a K factor which takes your your performance and and like gives you a penalty for the entire for the entire EFA blocks of four hours.

Got my fingers all wrong there.

Also shout out to K factors first time they've appeared on the podcast.

Oh really? Oh my gosh.

Think so.

Well that's that's pretty impressive. So I mean there's a whole they are very well designed markets and then lo and behold, we find out after a few years that there is all sorts of kind of misdelivery going on but they're not. They were penalizing some of it but not all of One example of this is reserve energy volume. So in order to make sure if there's a fault on the system the frequency and the frequency response can be delivered, you have to hold your state of charge within a certain bounds and if you contract more the bounds are tighter, you contract leads, the bounds wider. That that bound offers up you some flexibility in your state of charge. When you can trade volume in that, right, you can contract a little bit on frequency response and then actually buy and sell and try and generate profit.

But they weren't enforcing the response energy volumes.

And so you could see that providers were just trading, openly trading, getting really close towards top and bottom of state of charge by basic modeling, you could see that. And so what what you had, if you don't police things in a market that's open like this, especially frustrating when you've designed it so well, traders are gonna trade, people are gonna try and make as much money as possible and the market will respond to how much you penalize. Now, National Grid and ESO again, I'll give them give them kind of some leeway here because they also move these markets out really quickly. So it's a real, you know, it is a real balance and there's a huge consultation out at the moment. So these are getting stricter, they are gonna police them, they are bringing it in now.

But a lot of the tech has been built, a lot of the investment into the accuracy, reliability of the tech is built. So it might take a bit of time for those services to be delivered as well as they could be. And the important point there from my perspective is the better you deliver the services, the less you need to procure, the less it costs the consumer. So, it's like, it's kind of like a virtuous circle of making sure that these things are delivered correctly generates value and makes the whole system work a little bit better. So that would be that would be the one thing I'd like to see a little better and it's coming which is great.

It's more some more clarity around the rules and adherence to the rules in in the processes because it because it means that from a CTO's perspective, it gets much more clear like what you're trying to build. Yeah. Yeah. Okay. Okay. I'm with you then.

The one thing that the industry can do and I and it part they have put provision in this and I understand why this is complicated, maybe this should be my contrarian view is publish the performance data for each individual assets.

I I really agree. I agree with that. Would love to see data on penalties that were allocated. I think it would make it so much easier to people for people to understand really what's happening in the market. Yeah. There are I think reasons why Nito would struggle with that.

I and I understand that's that's that's an issue. But if you imagine that the analysis you guys could do and then this question of who's a good optimizer or not is like it adds so much color to the tech but also the hardware, right? Because if assets have to adhere to a certain accuracy level but it's not policed and you as an asset owner have got two decision points of, well am I gonna buy this one or this one? This one's way more accurate but actually the market doesn't really penalize inaccuracy, you're gonna go for the For the less accurate one.

And I think we so a few reflections from the Moto side, we we definitely saw when penalties were being enforced that behavior improved massively. And so I think there was almost like a a sort of squeaky wheel a little bit to kind of people were nervous about it being enforced. But when it did get enforced, strategies changed and people were more cautious. And then, actually, we had better delivery of the service.

So I would I would say that, you know, penalties work Absolutely. For sure. I think from from our side in terms of tracking things like penalties, we can build the logic to say when a penalty should have been applied. But if we're not sure that a penalty has been applied, it's then really hard to to to sort of put it onto an index and to say that it has actually happened.

Yeah. And so as soon as it's transparent, we can do it. And if you take a step back from all of this, like, why is anyone here doing this? We're doing it to try and build a better system at lower cost to consumer and leave gas in the ground.

That's that's really where it all comes from. And I think when you talk to Nisa, they're very much aligned on that. And the industry is very much aligned on it. So perhaps it goes back to your earlier point, which is around sort of that transparency and trust. The more that they can give that transparency and trust, then the more that we can do good stuff with it Yeah. To make the industry go faster.

I think there's there's another point here and this is the whole digitalization piece like we've started a high functioning software company in the last ten years. We've built everything from new and even within our within our tech, we have some things that are considered tech debt. So, there's things that I wish I'd have made decisions differently and that's just from ten years. So, the whole digitalization of the energy industry is dealing with legacy systems which basically just aren't set up for the sheer amount of data I said earlier on, like seven billion data points a day and we've only got you know thirty, thirty five, forty assets on the system.

So, trying to deal with that at the same time it's like what's the analogy trying to change the change the wings while the airplanes in the air. I have so much sympathy for NISO and the progress they've made but the entire energy industry and you know, again coming back to Renko and why we exist is we believe we can build, we can bring that modern technology, that modern data management approach to large energy companies that that is our that's our reason for being and therefore make them more efficient, therefore make them invest in more renewables and and flex.

And there's an exciting thing in here, right, that we're we're talking about this in GB as it's happened. But also there's a global context to this, which is so many system operators are just coming up to this. Right? So I certainly see this as being a very like, hopefully, this is a really interesting thing to listen to for someone who's maybe outside of GB as well as someone in GB.

Yeah. I wanna come back to one thing you said. So moving on to the tech debt side, maybe that's bit of a leading question. What are some of the keystone decisions you've made as CTO that future CTOs will have to live with?

The the key decisions I've made is the way we've structured the team and the people I've hired. Now, software, my background's in energy engineering, before that all sorts of other stuff that you can go on my LinkedIn and see. A band. Including professional singer in Japan which but the key thing there is understanding the problem set as a CTO of a company developing software is to hire and build a high functioning software team.

But also building people, one of the things again about the energy industry is such a core tenant for the future of this planet and how we live on this planet and how we how we manage our energy. We're attracting people from all other industries and the top tier of talent. In software we talk about T shaped engineers and usually that means you've got one kind of deep specialty but then you're across front end, back end, some different languages and maybe some infrastructure. What we get the team at Orenco I see is you get this specialty, you get the T shaped traditional but then you have this capital T of engineers who understand the battery behaviors or the markets or the balancing mechanism as well as anybody because they are in the detail of how it all works.

And I'm developing that and seeing that grow and developing that into the team. I think any CTO who would come in after me would be very, very happy with the team I've built. Okay. Because it's just such an enthusiastic, passionate, dedicated but just talented group of people who are able to like manage the crazy changes that come kind of almost on a monthly basis and do it all still while having fun.

I love that. And then let's talk about maybe asking that question in a slightly different way. So what is one thing that you now know that you wish you knew five years ago?

There's probably a host of prioritization calls that I would have changed in terms of when we schedule different things or whether we built something at all.

But all those decisions are kind of part of the journey philosophically. I think that's kind of part of it, you can't really you can't really take out any of the big humdinger mistakes because then they they led to you being more robust in decision making. I think the one thing would be a bit more of a personal message to myself.

I'm a human like lots of people who live with kind of impostor syndrome and yeah, insecurities that cause me to burn way more energy than I need to. And I think what I'd love to do is just go back to myself and just be like you're doing alright, like just crack on. You're you're you're you know, there's you're doing things that nobody else has ever done before, you're never gonna get it perfect. So, it's a it might be a bit of a cop out but I think that's what I'd like to and I guess anybody listening who has that mindset and sees Renko as a great company, like that's the reality is that why I deal with the same things that loads of other people deal with and I think it's crack on, get things done, don't worry too much about it and the mistakes are as important as the wins.

Okay.

I like to think we've given people a a spring in their step as they're kind of on their way to work with that with that thought.

Good.

Okay. I'm gonna wrap up with two final questions. So first of those, is there anything you'd like to plug?

We've talked about the evolution of Arenco and I think one of the things we have been around for a long time and I think we've discussed it a fair amount, but it's communicating what we are as a company. We are a technology company who have this platform Nimbus, which can support across scaling large portfolios and deal with multiple kind of problem sets across asset management, data management and the actual optimization and trading. And I think people still potentially see us as an optimizer, that's one part of one of our tooling allows us to be an optimizer. But the fundamental is a really modern technology company that can help with digitalization at a at a global portfolio scale.

Okay.

That's kind of what the message I've I I'd love to get back out there to people.

I like it.

And then moving on to what is a contrarian view that you hold which could be a bit box office.

I'm not sure I've got a box offy box offy one.

Okay, yeah, let's let's try this one. So, I think the way we've we've alluded to it a little bit, the way the energy industry operates and particularly consults and changes, I think can evolve and it can use it can use the experience from other industries to do this and it's all around the digitalization. And one of my main frustration is working groups and steering groups that I sat around working on PowerPoints and maybe maybe a work, maybe a whiteboard with post it notes. That works to a point, but the industry we work in is so technical. I think we move need to move towards a kind of rules as code, infrastructure as code using open source approaches to solve problems with engineers and product teams in the room. You always got our product teams to steer those engineers because they will go crazy. But you've got an industry full of incredibly innovative companies all working and building tech at an incredible rate driven by AI, building it at even faster rate.

Orchestrating that into open source methodology to get agreement with industry and this is, you know, something I'm working with NISO at the moment to try and bring in all these rules around the markets, these really complicated markets, how they operate in between. If you can get the the equations into code, share those, get agreement, they get tested in the environments of all the technologies, they push push the barriers, see where all those just slightly out there instances where current code base falls down and improve it. You then can hit markets, launch markets, launch new concepts with technology that's already been tested and agreed with industry. And I think they, there's an incredible opportunity there with the amount that's been invested in this industry, the intelligence that's out there to really accelerate particularly for GB and then be a market leader on basically how to operate really complicated future state energy systems.

Open source system operation is a hundred percent box office, no less than I expected.

Roger, thank you very much for coming on. You've been a fantastic guest. I think people will have got this fantastic insight into how the whole optimization asset management works and yeah, I've learned a huge deal.

So thank you for coming on.

Brilliant. Thank you for having me. It's been great.