Intelligent Decarbonisation - A Chat With Prof Markus Kraft

emissions. In this climate 21 podcast, I will showcase best practices and thought leadership by SAP, by our customers, by our partners and by our competitors if they're game in climate emissions reductions. Don't forget to subscribe to this podcast in your podcast app of choice to be sure you don't miss any episodes. Hi, everyone. Welcome to the Climate 21 podcast.

My name is Tom Raftery with SAP and with me on the show today I have my special guest, Markus. Markus, would you like to introduce yourself?

cool. So the first thing you can do is to use low carbon or carbon neutral fuels. So we do some fuel research that investigate this, and a low carbon fuel could be a biofuel and the various versions of it, it could be a synthetic fuel okay, for example, it could be hydrogen, it could be methanol, it could be a biofuel made of algae, it could be sort of a mix between the two. So what type of fuel is the fuel of the future?

And here, the ultimate question is, can we produce a low carbon fuel at cost that is competitive with diesel or petrol? That is the key challenge. And that is a super hard problem. Okay, and of course, do we have enough resources to produce these fuels on scale? Right, because if you have biofuel, depending on what type of biofuel it is, you may eat into our food crops and that is something we definitely don't want to do. So you have to be

clever about it. So, and another point is not only how you make these fuels, but what is the best fuel in terms of if you still want to burn it. Like for example, in a jet engine, you know, an airplane which probably will be run on a hydrocarbon fuel, a low carbon, hydrocarbon fuel, whatever that means, that's a hydrocarbon fuel with this wonderful energy density that is not based on fossil sources. That's the key point, okay. And that is this is probably the technology of the

future. as far as air traffic is concerned, and here, we need to ask a question, can we burn such a fuel cleanly? So are there any emissions coming out, okay, because that in turn has a great impact. So we also do some research on emissions in our programme. Okay, so that's the first point. But I told you there are three, okay. I will be a bit shorter in the next two. So another one is to capture carbon, okay, to separate it from the exhaust, and then do

something with it. You can either then store it or utilise it, which brings me then back to the synthetic fuels. So we have one programme where we take CO2 and try to turn it into either fuels or high quality products that sell for higher price, okay. But this is in terms of technology readiness levels, at quite a low level, which is a point we can discuss, and is very important. The third point brings me to increasing efficiency of how we use energy

at the moment. And this, in turn, brings me to digitalisation. And the big question here is, can we use digitalisation and artificial intelligence to reduce or to increase the efficiency with which we use our energy? Right? So these are, at least in our book, the three challenges and we have in our big research programmes, sub-programmes that work on each of these streams..

usage of the fuel? I know, some people have looked into potentially using hydrogen for planes, but I suspect from what I've read, and please correct me if you found otherwise, but I suspect hydrogen is a non runner for planes and we'll end up using some form of synthetic fuels as you as you talked about earlier, feel free to jump in and tell me that's rubbish or, or what your what your findings are for the research into fuels.

not, okay? Of course, if you think about hydrogen, hydrogen is best transported if it's connected to a carbon atom. Okay, hydrogen is not a good molecule, it is very small. It doesnt have a high energy density, if you store it, it's gas, it's a gas, it can, you know, it's hard to transport.

energy density. If you think about that, you know, a person like me, you can't see me now but I take you I tell you that if I was given food for four weeks, I would be still alive, okay. And this is only because of the high energy density. Now if you look at the robots in, say, for example, the leading robot labs in the country, all over the world, then you see, they're all connected with a cable to the plug socket. And the reason for that is the batteries, you know, would be,

they wouldn't last a day. And your body keeps all the energy constantly with you. If you don't, you know, there are many people who don't eat for a whole week because they're fasting. And it's not a problem, they feel even better. Sometimes. that is what they say, I'm not recommending this now to you. But what I'm trying to say is that hydrocarbon fuels have a very high energy density. And that is beautiful. They also come in liquid forms, which means it can be very easily

transported. This is the reason why if you go to a petrol station, you can fill up your tank so quickly. And this is one of the problems with electric cars, where you have the charging problem of batteries, okay? And we need to, we need to address this problem. There are many reasons why you should use an electric car. But the speed of charging is not one of them.

idea. And I don't want to overdo the hydrocarbon fuel argument too much but what Im trying to say is it is a very important product which we also need for our chemical industry to produce the materials that we are all using and being able to produce such molecules without fossil fuel, that will be an important skill which we need for the future, its absolutely clear.

domain. But now, and this is already, I told you about digitalisation, this is where already the bridge comes in, okay, because of artificial intelligence, we can now represent chemistry and materials already in, sort of, the computer. So, automating a laboratory, automating, creating a world model this is what we are actually doing that will help not only with trying to make, for example, a district heating system more efficient, it also helps us to accelerate

the process of science. Okay, so you can have a lab robot for fast screening of different chemicals and reactions, so you get fewer products you don't want, you get fewer wastes, you get higher efficiency, you make the process cheaper, you use less energy. And that all because the whole process has now been automated to use optimisation, machine learning and all the mathematics that is known to, to accelerate the materials research and also to

do the process research. And I believe that this, now we call it world model, that this attempt of making things virtual will help us to accelerate the overall progress of our research. And hopefully, that will give us a virtuous circle, virtuous circle where we produce something better. And because we have something better, we can be faster and so on. Yeah, and you have seen this, for example, when you look at the cost of solar cells, okay, where we have better materials, cheaper, and

so on, okay. So this is a, an important aspect of our research. And the world model I've been talking about is called The World Avatar project. And

the problems, so, you know, imagine you have a digital twin of your car, you have a digital twin of your fridge, you know, and you have a digital twin of maybe your, your stereo. I don't even know whether that still exists, nobody has a stereo, that tells you something about how old I am, okay. So, but these digital twins need to talk to one another. Okay, so you get into what we call an

interoperability problem. Okay, so we have developed research that addresses this problem and allows digital twins to talk to one another. So in the first instance, we try to get all the components of an industrial chemical park here in Singapore, that's Jurong Island, and that led to the J-Park Simulator. And we then use this concept further and build then The World Avatar

project. And that project allows basically the inclusion of everything, because you also, you use technologies called knowledge graph technology that has been developed for the last 20 years. And that allows you to completely distribute your data, to have it interoperable, and to make the digital twins talk to one another.

to you. And then you have senses, and that can be the temperature on your skin, it can be what you see or what you hear, okay. So you have also a virtual representation of some parts of your body, so you may have heard that if people lose an arm that they still have pain in the arm, although the arm is not there anymore, which tells you that there is a digital twin of that arm in your world model. And the same is true for our

laboratory. Okay, so we have these sensors, so, and now, we also have a cabinet with flammable liquids, they have RFID tabs, okay, so if at six o'clock, not all the bottles have been returned into the flammable liquid cabinet because one student forgot to put it back, there will be an email to our lab person who would then follow this up, okay, this is a safety measure, but it could also be a fume hood that isnt

closed. So there would come, oh, fume hood XYZ, okay, I know, I've noticed I'm not closed, then there is a process, an agent that would look at that state, and would tell me, oh, the fume hood is not closed, let's go contact somebody who can close it for me, right. And of course, for you as a person similar, okay, you haven't had

some water to drink. And after a while this, you know, this feeling that I am thirsty gets stronger and stronger and to the extent that you will do something about it, and then you go and drink something, until you get hungry, and then you eat something, okay, and then you get tired, and so on. I have to say that you may already have noticed that I'm an enthusiast and that I like this work very much. And the inspiration for this, I draw from Marvin Minskys book The Emotion

Machine. So the listeners, if they are interested in one of the fathers of AI's greatest works, then they should read this book, I think it was published early 90s. And this whole idea of systems of agents that that basically, form a society of mind, is described in this book beautifully. It's

really an eye-opening story. And of course, at the time when this book was written, the technical conditions werent there, hadn't been made to be able to implement it, but now, now we can experiment with it, and if you combine these ideas with machine learning, then you get a very powerful system.

technology to roll out okay. So,

this is one aspect that you can use, but the second aspect is now, we are talking in the UK with the company that administers the smart meters. So now, of course, you can represent every house every consumer in this world model. And of course, you can then say, right, how do I manage my energy, and if you have a car, and you don't use it, and you want to use your car as an energy storage to the grid, I may now develop a digital twin of your car that is able to trade energy directly with other

sources. So, we can look into different ways of how to deal with with electricity network and roll out and introduce payment mechanisms, new payment mechanisms that sort of create a non regulated market in the sense that the that the market always optimally reacts in an optimal low carbon way. Right. So that would be that is one research project that we are doing together with Nanyang Technological University here in Singapore.

good chemist. You need somebody who can work in all of these worlds. So we spend a lot of time training people to get them to the point where they actually make original contributions. So it's, it's very fascinating, it's very interesting, but it's also very challenging.

mitigation. So if you have extreme weather event, imagine, okay, now everything is digitalisation, you have senses of the waterways, you know, along a river, you have some heavy rainfall or something, you build up a wave, you can sort of predict how that would impact we're looking at Norfolk, in particular Kings Lynn in the UK , you can sort of predict how

this flood waves come in, or maybe a storm from the sea. And so with that we have, we can then look into the infrastructure, we know which of the infrastructure is vulnerable, we can then switch it off to avoid bigger damage, we can have mitigation schemes, okay. All online. And this is the first time where our Knowledge Graph concept is not only academic research, but it's really implemented. And this is I find this terribly exciting.

big theme, okay, digitalisation and decarbonisation are two main, big streams that keep us occupied. If you look at governments research programmes, you will find one on

decarbonisation and you will find one on digitalisation all over the world. And the question is, how do these fields, you know, connect to one another? So in the book, we contacted people from academia, but also from different types of industries and NGOs. We also have about ten interviews with leaders in the field. So these are presidents from universities, leaders of international funding organisations, we have people in, with NGOs, Grow Asia, for example, so food agriculture

also covered. We have people from industry, Siemens, for example. We have consulting people, cybersecurity, so we the book is trying to cover pretty much any aspect of the, of this greater topic, both by scientific articles, and these interviews, to give you different perspectives on this. And I found it, to be honest, it was a fantastic experience to connect to all these people, gather data together, and it will now come out right before Christmas, which brings me back

to, you go and order. But you will get it through the university libraries. And so it's relatively, should be not too difficult to get subscriptions. And the chapters that I have written are also there in preprint format. And so that, that's a good thing. Its a book by Springer. And as I said, this combination, artificial intelligence, decarbonisation comes in many shapes or forms. One thing we also address is I already mentioned cyber, cyber threats. We also look at the existential

risks of AI, okay. So I don't know whether you've heard of Nick Bostrom, he had this, wrote this book Superintelligence where he talks about existential risks. And in my view, when we, when we do something like developing a world model, weve really got to be a bit careful that that we do it in the right way, okay, and that

the book. And here they concentrate specifically on the use of artificial intelligence algorithms on climate prediction. Right, you know, but there are also other aspects, where people try to use the machine learning aspect to look at this specific, specific problem. Okay.

technology can be. Right? So this bad person uses the DNA technology to be a bad person. So that's all I can say. Okay.

than before? And the answer must be, we are not using any fossil fuel, okay. Only then then and only then we are making progress. Okay. And this is going to be super hard, because

I told you earlier, fossil fuel is too good to be true, it has super nice properties and has been, gone through the biorefinery of nature for free. It is very hard to beat on cost, which basically means that if you want to be successful, we need policies that change the

economics of our processes. And if you ask me, where the most important point is that we need to address in order to make progress, then I would say it is not a technological, it is policymaking one we need to address to make fossil fuel too expensive to be used. That, in turn, will have strong impact on the social fabric of our societies. We need to understand how to mitigate the negative financial effects that increasing a carbon tax will have on the poor people in the

society. And I think this is the key point I want to make to talk to your politician and say, we need to solve this problem.

change that. So I am now going to apply for Spanish citizenship, so that I can get a vote so that I can cast it towards candidates who have a strong climate agenda. The only reason the only reason I'm getting Spanish citizenship, it's so that I have a say, and that I can vote for people who want to make a climate difference.

published. Also, please don't forget to rate and review the podcast it really does help new people to find the show. Thanks. Catch you all next time.