NXP’s Mark Swinburn on IoT and AI Driving the Future of Energy Management

Mark: And we need to find a way of managing, in a safe way how we can protect our grid but also manage the energy that flows around the grid. And then we need to think about energy security, With

geopolitics as they are today, we need to think about how we. Manage our own energy in each region.

And that can be by improving the growing reliance on renewables, for example.

Because we could easily predict that the cost of fossil fuels will increase in cost and the energy cost can, normally be used as an indicator for a country's prosperity.

And it's also a determinant for economic competitiveness.

Sure. And that also has an effect on investments. Let's go to the extreme. If your electricity was free, then you could imagine investment would be significant.

Yeah.

Because one of the big costs for, any economic activity is, energy.

So, energy cost and energy security are fundamental to a nation's prosperity.

And then the final challenge then that we should consider is, around customer engagement. And many of the energy suppliers that I talk to, one of their biggest challenges is how do they. Engage their customers. How do they become more transparent with their customers? How do they get their customers to help them with the energy management as well?

So that's a significant challenge

that

we hear from, a number of different stakeholders. So that's the kind of the headline challenges, yeah.

Ruth: To me it also seems that there are so many contradictions in all these topics you just mentioned. Like for example, with the energy grid, the contradiction is at one point it needs to be resilient, and we need to have so much energy stored to keep up with all the business challenges we are facing.

Yet at the same time, some windmills, for example. Produce too much energy and then it can't be stored, and the energy is wasted. So, you know which contradictions I'm talking about and at the same time, absolutely everybody's talking about yes. Climate change, the challenges we are facing. Yet at the same time, AI data centres, server farms using or consuming more power than ever before.

So, to me it's interesting, like how do we solve these issues when they're contradicting themselves?

Mark: you're absolutely right as. We evolve in energy management, onboarding, renewables are the answer. Okay. to rising costs. Rising demands. But we need to have a way of not only creating but also storing.

And if we want to generate and store, then we need to manage. And that's really where all of the, current activity is, aimed at. This is about bringing on new renewables or new resources to, to manage the capacity demand. But at the same time, being able to store and distribute that energy more efficiently, so One of the biggest growth areas actually in energy management and, energy in general is energy storage systems. So, we are seeing significant investment in solar, in wind. But now the big investment really is in energy storage systems. Because as you say, we can create energy from solar during the day.

We can use, other energy forms, but we need to store it. Because quite often we create more energy than we can actually consume, or we need to store that. So that's a big area of investment, for, not just us, but on a national level. But also, we see that in homes, people are now investing in solar panels for the roof.

The advent now of electric vehicle means that potentially you can use the battery. Within the vehicle as an energy store. And you can transfer the energy to the vehicle when the electricity is cheap, and then you can store that energy from the car back to the home when the local energy cost is high.

So, we are seeing storage as really one of the big key changes in the energy system as, we move into a more digital system where everything can talk to everything. And now we can start managing either from a, at a sort of a residential home. Scale all the way up to, large building, large premise regional, and then national areas as well.

And that does bring into to a question, how do we manage the systems? First thing we need to do to bring up that efficiency is to look at. Optimization. So where are we consuming? Where are we generating, where are we storing? And that system needs to be as near real time as possible, and that is really pushing a lot of the technology to the edge.

Okay. Okay. Traditionally, we may have thought of a lot of these management systems and a lot of the control is in the cloud or in the backend. Yeah. actually, what we're seeing because of this, near real time, requirement, we're seeing now more and more intelligence is being pushed to the edge.

And that also has an advantage because now. Their management is at the edge. It's a lot closer to the end customer. And we mentioned earlier on, didn't we, that one of the big challenges was customer engagement. So now we can almost connect the customer directly to the management system.

Okay. To interact, to change behaviours, or at least be informed. On behaviours. So that's one of the benefits of, being certainly at the edge

Ruth: is the infrastructure ready for that? Because it's, it sounds like there needs to be a lot of silos to be broken down so that the energy management can be dynamic and intelligent.

Mark: Yes. one of the biggest challenges we've got is they don't necessarily communicate. To each other. Okay.

Then that, that calls into questions, how do we have a common language? We and our customers are all putting in that intelligence at the edge. We're making those systems more intelligent.

We're adding management to these energy resources, but they have to talk to each other. and that then comes down to local legislation. Or indeed, even on an EU level, we should have a common language. And a common way then to, for these, for this equipment. To talk to each other and have the same specifications and requirements so that we can get interoperability.

So, whether it's a, it's an energy storage device in Germany, or it's in the Northern Hebrides in the UK, they should all have a common way of communicating and have a common behaviour that we can, we can use map and take to our advantage.

Ruth: So, what does this shift look like? And maybe you can share some, real life uses cases as well.

Is this something that is already happening or something that is just planned for the future?

Mark: Yeah, the shift is, very much here today and it's been driven by two key challenges that we see, in, the marketplace. The first one is. For, any, efficient control system, we need to create a rapid response to any quickly changing situation.

So that's fundamental. and as a result, we need to directly connect these inputs and the sensors at the actual source. So that's a new trend that we're seeing more and more, and our customers are behaving in that way. Rolling out. New equipment to manage that, real time response requirement.

And the second thing is a second challenge rather, is that we need to build intelligence into those resources, taking in that input data and making intelligent decisions in an autonomous way. And that could be simply around, load balancing. Perhaps we want to load shifts, so let's delay consumption until a later time.

We may consider energy storage as opposed to consume.

Or as well as taking the energy perhaps from your vehicle and then using that as an energy generator to your home. And that really affects things like distributed energy, resources like your EV charger, maybe your, energy storage device, your solar PV on your roof, and indeed wind sources as well.

And they all fall under this category of, distributed energy resources. So, when we put those two factors together. It dictates that we have our intelligence at the edge because we need to be real time. And because we are, at the edge, we can actually filter out some of the data, some of the noise that's not actually relevant, and then just focus on the data that then drives those, those decisions.

So, we have this. New creation of a distributed energy resource manager or a doom for, short. and that's incorporating the AI and machine learning analytics, and that helps to make, decisions autonomously. And that in turn then ensures that we have an optimized and efficient asset and also that it, it behaves in a reliable and safe way, in a timely manner.

So that's how we see a shift today. And that's where a lot of, our work and activities are at the moment.

Ruth: It sounds promising. You mentioned AI and machine learning as a key driver for the intelligence at the edge. Can you elaborate on that? What benefits does this offer?

Mark: Yeah, absolutely. AI is fundamental now to, to what we do particularly at the edge because it gives us three very potent insights and those are namely detection.

So, we can. Consider anomaly detections such as temperature, for example, we can look at peak power, we can look at noise. There may be arc, and electrical faults or indeed other safety issues. And also, we can use detection for potential revenue loss. So, AI is giving us insight that we didn't have before, but it also allows us to now start to make predictions.

And those predictions can be around, consumption and generation patterns so that we are making sure that the grid is. Creating the energy when we need it but also storing it when we have access or access. But also, we have climate effect, which will have a direct influence on how energy is created.

But we can also use predictions around maintenance. So routine maintenance is a common use case behind, machine learning. also, we can use it for, predicting the state of health of our system, and that could be the state of an energy store, but it could also be, the state of health of a grid.

I. And all of this then can be improved through continuous learning, specifically with machine learning techniques. So those are the insights, and then we can take those insights and use them in, real world. So, for example, we can use those AI analytics to create autonomous decision-making systems and that, will drive up the efficiency because.

If we can maximize utilization of our assets, then we can drive down the cost and we can manage the energy in a more efficient way. So again, we talked about load balancing, load shifting, whether we store or export.

And we can add safety to that. And the final, I guess the other benefit we get from AI machine learning is.

We can actually use the decision making from the artificial intelligence to improve the customer's interaction as well. So, we can propose, or the system can propose better services for the consumer. But also, insights into the way they use energy. Perhaps how they're attached devices or appliances are behaving, and that all could come together to actually reduce their electricity bill.

So, there are some, clear advantages of having AI at the edge and, close to the consumer.

Ruth: It does sound actually really easy yet, there are so many barriers and hurdles, you mentioned some of them already, like investment. I suppose expertise is one of them. And then standardization, as you said, there should be one unified language so that within Europe it would be easier to exchange power.

For example, if I don't know, countries. Along the Northern Sea produce more wind power, and countries in the south produce more solar power and then it would be nice to interchange, for example, in the wintertime or vice versa. Can you share some strategies or partnerships? Do you have a recommendation?

Mark: Yeah. That some of the barriers, so yeah, I think it, it's fundamental that governments, companies and investors all together need

to get behind the, this clean energy transition and stop hindering because having a common language and a common approach, will, certainly speed up the rollout and the integration because.

Uncertainty is a real problem. It's one of the key decision makers for investment. So, where there is uncertainty, typically there's a lack of investment. So, I think governments and companies, can, help to improve a situation, reduce that risk with, standardization, for example.

There are, significant benefits behind having this digitalization and energy management system around jobs, for example, there'll be greater energy security.

We talked about that earlier on. But we also have environmental benefits, so cleaner air, for example, and a safer climate for everybody. So, there's clearly. Real benefits behind doing this. Both, e economic benefits as, as well as climatic. and the advice from NXP would be around, building partnerships.

Building the right partnerships within industries is fundamental, and that has to involve adopting, proven AI or machine learning technology. So, at NXP we have. Acceleration of NPUs across most of our processing platforms but also build scalability. So have systems that can be deployed at the edge in a very low cost, low energy way, and then scaling that performance requirement up throughout the different levels of, the system, of the grid so that you have this compatibility that we need.

Customers should always consider building third party ecosystem because. It's unlikely that not one company or one individual knows everything. I think it's it. We all appreciate that. So having a strong third-party ecosystem with partners that can bridge gaps, certainly around knowledge, and skills is fundamental to what we need to do to make sure we get the right kind of support, the right kind of training, the right kind of education, because.

Machine learning and AI still relatively new and not everybody has mastered it. And we would be relying on, third parties and, partnerships to, to bridge those, gaps.

And also, standardization is critical to this, if companies are able to join some of those governance committees, they're at the national and international level.

It's really important. And this really should be seen as a necessity and not, not a chore, which it can be. And then we need to look, from, companies to when we engage that they are able to, to see all of the investment opportunities. That are available to them. So, in Europe, for example, there are a number of directives that actually will, provide money and funding

To help the company or, the regions to move into this, digitalized grid, and energy management system as well. So, there's a few things that, that we can all do. But also, when we engage with our customers, what we recommend. Sounds great and great advice.

Ruth: Thank you. I suppose this brings us now to the topic of security and resilience.

Since cyber threats and the energy systems are on the rise, how is NXP integrating security measures into its technologies?

Mark: Yeah, it's, clear and, we all appreciate that. This, security or cybersecurity landscape continues to evolve. In particular when we apply it to the grid and energy management as, we connect more and more products together, the actual attack surfaces for this potential cyber threat.

Is continuing to expand. when we talked earlier about distributed energy resources, advanced metering infrastructure, that's now creating multiple entry points to malicious actors because we have different, components, different products coming from different vendors, different suppliers, and each of that could be opening up, an entry point for, for these malicious actors.

Also, it's worth noting that any an energy management system. Should be considered as part of a, nation's critical, infrastructure. it means that it's, become A target. Even more than, let's say, our daily threat actually being part of a nation's cri critical infrastructure, actually increases the exposure and the intent.

And what we are seeing is different threats coming from not just, state organized groups, but also groups that are aligned now with states also including, terrorist gangs. And we were talking significant threats here where. Organized criminal gangs now are using things like ransomware, and what they've actually done is they've, actually reduced the entry barrier or the cost for people now to, to begin hacking and extorting money from different organizations.

So, it's, I. It's quite significant. it's, it continues to, evolve and become more sophisticated. When we think about ai, we talked about all the positive benefits of AI and machine learning. Actually, you can then also use those in your favour, in a malicious way. So, we need to be. Aware of these new challenges that are coming.

And there are attacks directed specifically at AI algorithms that we need to be aware of. And that's before we think about, okay, what about quantum computing? What does that do? So, it's clear that what we have continues to evolve, and it is not a stationary picture, but at NXP, security certainly is at the heart of what we do.

We very much, approach. Technology, integrating technology, and we have two very distinct approaches. One is where our security is onboarding the processing device, whether that's, a microprocessor or an application processor where we have dedicated security hardware that's responsible for providing security services to, the device in an autonomous way.

So, it's not affected by, Potential breaches within the device. And then we also have the second half where we have standalone security. These are high security modules. They are what you might associate with either enterprise security or bank grade security. And again, our customers can actually integrate that technology into this system to bring things like root of trust, which affects, resilience, secure, boot secure over the air.

So, there's a number of functions that once you've got an established route of trust, that you can then start building security services around to try and improve your resistance, but also your resilience to, security. And that's, fundamental to what we at, NXP do. So hopefully we can, we can alleviate some.

Of the concerns of, of our customers.

Ruth: Fantastic. Yeah, it's always, difficult Nowadays the pace has increased so rapidly that it's getting more and more difficult to, stay ahead of, absolutely. A security issues. It is unfortunately a shame that whenever new technology emerges, there's always the good you can do with it and.

And the bad does

Mark: two sides. Yeah.

Ruth: Yeah. Usually follows really quickly. That's a shame actually about Yeah. How it is, right?

Mark: one thing, I read a report recently that, that said in the last 12 months, over 40% of national infrastructure companies had suffered some form of breach. But that's all of the energy, whether that's water, electricity, and gas.

So, when you put them all together under what we might consider critical infrastructure, yeah. Over 40%. Had suffered breaches. the area that we work in, energy Is critical. But it's, also a natural target for large organizations that we see, active every day.

Ruth: Then there's obviously also just a, if it's not criminal intent, sometimes mistake happens.

I think it's already two years ago I was in London for a rugby game, and on our flight back, actually, the whole British flight security system. In all of UK just shut down because

Mark: Yep. Somebody it shut down. I remember that.

Ruth: You remember that?

Mark: yeah, we all remember that. Yeah.

Ruth: I think

It was one cell in an Excel file that was wrong and that brought the whole system down.

Mark: It was that simple, wasn't it?

Yeah. Yeah. So, this comes down to, testing. Okay. And interoperability. So yeah, there was that. it just goes to show that, you could, if you can find a single-entry point, and it might seem quite mundane. You can have a significant cascading effect.

So, what should have been a simple, file, not find or error. Undetermined actually, cascaded through the whole IT system and then shut it down. I.

Yeah. we need to understand potentially where and how attacks are launched and then try and build, safety valves in to prevent things like this.

Ruth: We are almost at the end of our show. Yeah. If you had to pick one emerging trend that will redefine energy management in the next. Five to 10 years, what would it be?

Mark: Yeah. I think what we are seeing, certainly as the trend is that we are now moving towards this autonomous, secure, and real time energy system.

We know already that the AI is fundamental to that transition. So, for us it's about developing new tools, new models, and being able to deploy that in a more simplified way but also maintaining those models in the field. I think the trend really is going to be around developing and launching models to, to improve efficiencies around real time and consumption.

So yeah, ai, for, me and for what we do is, the real focus as the new trend I. Terrific. Is there anything I have not asked you that you wish I had asked? So, there is another area of, industry that is also emerging. It's called containerization. So, containerization Is where you are able to run functions on a device. In an isolated way. So, you can, for example, if you have a product in the field and you want to add a function, a temporary function, you can give that function, download that function to a device, it runs in a container, it will complete its task. And then potentially you can update it, remove it, for example.

But the container itself allows you to manage. That function much easier. So traditionally, you may have to update the complete firmware of a device if you wanted to change its functionality. But now using containerization, you can actually update individual functions. It's almost like your app on the phone.

You can install a new app to do something, and then when you're tired of it or you want to change it, you can delete it. And that is coming to even small MCUs that, that exist now at the edge. So that flexibility that we didn't have before is coming. and containerization does that in a very secure and organized way.

So, I wanted to mention that because I. I think it does solve a number of problems that we get, certainly in terms of maintaining a product over its lifetime in terms of adaptability and functionality. So, containerization absolutely is one worth mentioning, I think.

Ruth: Fantastic. Yeah, I was just thinking that this should, speed up deployment cycles quite a bit. Yes. If you had to pick a song for the soundtrack of the episode, what would you choose?

Mark: If, we assume then that, that, the people that are listening in will have an interest in energy, because today we've been talking very much about energy and energy management, then I think we all will have the ability to enable this transition. I think that will then give us, the onus is now on us. So, I think a good song might be, snap, and the song is the Power because that's what we all have, the ability to change and get involved. So yeah, snap and the power would be, I. I think it would be a good start.

Ruth: Lovely. It will be a great, addition to our compilation. Thank you so much.

Mark: It's my pleasure.

Ruth: Fantastic. Thank you so much for all your brilliant insights today, mark. It has been terrific chatting with you, and I hope you enjoyed yourself.

Mark: It was great. Thanks for the opportunity and yeah, hopefully, we'll have. imparted some benefit, that, our customers and colleagues can use in the future.

Ruth: Absolutely. And it would be great also to have you on the show again, maybe in the future and discuss what became out of all the ideas we discussed today.

Mark: It would be absolutely my pleasure. For sure.