WEBVTT - Modeling the Energy Future: Ants, Birds and Dr. Strange 0:00:00.200 --> 0:00:02.440 So right now in my desk have got three books. 0:00:02.840 --> 0:00:06.200 The first one is by David Epstein called Range, How 0:00:06.280 --> 0:00:11.080 Generalists Triumph in a specialized world. Fantastic book. Can't recommend 0:00:11.080 --> 0:00:13.399 it enough. The second book, and at three hundred and 0:00:13.400 --> 0:00:15.760 seventy five page is I mean. Book is a copy 0:00:15.760 --> 0:00:18.880 of Banest's new Energy Outlook twenty twenty or NEO, our 0:00:18.960 --> 0:00:21.560 annual long term scenario analysis on the Future of the 0:00:21.640 --> 0:00:24.639 energy economy, just released in November. It's an Outlook all 0:00:24.640 --> 0:00:27.120 the way to twenty fifty, and his Beenest Flagship Report. 0:00:27.680 --> 0:00:29.720 The third book is called Kings of the Yukon, an 0:00:29.720 --> 0:00:33.080 Alaskan River Journey by Adam Weymouth. Here the author records 0:00:33.080 --> 0:00:35.360 his trip down the Yukon by canoe from Teslin Lake 0:00:35.400 --> 0:00:37.760 in Canada all the way through Alaska to where it 0:00:37.840 --> 0:00:40.040 lets out in the Bearing Sea. He's taking the trip 0:00:40.080 --> 0:00:42.760 to see firsthand the run of the Chinook or king salmon, 0:00:42.840 --> 0:00:44.360 and talk to the people that have been marking the 0:00:44.400 --> 0:00:46.919 time of their annual return for generations. But for me, 0:00:47.159 --> 0:00:48.920 the most fascinating parts of the books so far have 0:00:48.960 --> 0:00:51.199 been when he goes into detail about the behaviors of 0:00:51.280 --> 0:00:54.320 the fish, for example, how they know exactly how to 0:00:54.320 --> 0:00:56.880 get to their spawning grounds down to the tiniest tributary 0:00:56.920 --> 0:00:58.800 or creek of the river without ever having been there 0:00:58.840 --> 0:01:01.360 as adults. Well, today on the show, we're gonna keep 0:01:01.400 --> 0:01:04.080 going with the topic of animal behavior. Turns out the 0:01:04.160 --> 0:01:06.920 engine behind the analysis and NEO, the code that does 0:01:06.959 --> 0:01:09.760 the calculations and creates the scenarios in the outlook, was 0:01:09.800 --> 0:01:13.679 built after animal behaviors, specifically those of ants and birds. 0:01:14.160 --> 0:01:17.399 With us, we've got Ian berriman modeling analyst for PNF. 0:01:17.440 --> 0:01:18.800 He will tell us how the model, known as the 0:01:18.840 --> 0:01:21.920 New Energy Forecast Model or NEPHIM works and how he 0:01:21.920 --> 0:01:24.240 took cues from ants and birds and building it. We'll 0:01:24.280 --> 0:01:26.600 also talk quite a bit about Dr Strange from the 0:01:26.640 --> 0:01:31.120 Marvel movies. So Comberbatch fans get set. Okay. Our discussion 0:01:31.160 --> 0:01:34.400 is based on BNS New Energy Outlook. Being a users 0:01:34.440 --> 0:01:36.280 can get this report on BNF dot com, the BENF 0:01:36.360 --> 0:01:38.600 mobile app, and the Bloomberg terminal. As a reminder, BENIF 0:01:38.640 --> 0:01:40.760 to provide investment or strategy advice, and you can hear 0:01:40.760 --> 0:01:42.119 the full dist claimer at the end of the show. 0:01:42.200 --> 0:01:44.280 I'm Mark Taylor, and you're listening to switch on the 0:01:44.319 --> 0:01:55.200 benf podcast Ian Welcome, thanks for having me. So today 0:01:55.240 --> 0:01:57.600 we're going to talk about modeling. I used to be 0:01:57.640 --> 0:01:59.720 an energy analyst, and I'm not a modeler. You know, 0:02:00.040 --> 0:02:02.600 I'll be I'll be very blunt about that one. But 0:02:02.840 --> 0:02:05.200 for everybody that is listening that doesn't even know what 0:02:05.240 --> 0:02:07.200 a model is, can you just take us all the 0:02:07.200 --> 0:02:09.240 way back to the top and tell us what a 0:02:09.240 --> 0:02:11.600 model is and why they're important. It's a good question. 0:02:11.639 --> 0:02:15.000 So I think why models are important is because we 0:02:15.040 --> 0:02:18.600 don't always have time or space to try everything out. 0:02:18.919 --> 0:02:22.919 In reality, we don't have a physical way of testing 0:02:22.960 --> 0:02:26.840 this hypothesis, don't have the materials. When we're talking about 0:02:27.000 --> 0:02:29.800 models that look towards the future, we don't know what 0:02:29.840 --> 0:02:33.600 that is until it happens. So models help us answer 0:02:33.720 --> 0:02:37.639 questions which we wouldn't otherwise be able to answer. Okay, 0:02:37.639 --> 0:02:41.280 So it helps us get through more data and ask 0:02:41.360 --> 0:02:43.639 more stuff of data that we wouldn't otherwise be able 0:02:43.680 --> 0:02:46.799 to do. Exactly Okay, cool. I think it heard like 0:02:46.840 --> 0:02:49.560 could even be considered like like the model that of 0:02:49.600 --> 0:02:51.600 a building that an architect builds before they actually go 0:02:51.600 --> 0:02:53.280 out and build it, they can see what it looks like. 0:02:53.639 --> 0:02:55.960 You're good to see how it might work exactly. And 0:02:56.200 --> 0:02:59.240 related to that, there will be somewhere on that architect's computer, 0:02:59.400 --> 0:03:03.320 there will be computer model which has all the beams 0:03:03.360 --> 0:03:05.280 and all the weight in the building that is expected 0:03:05.280 --> 0:03:07.960 to carry. And there will be another model which simulates 0:03:08.000 --> 0:03:10.840 winds or earthquakes to make sure the building cap with 0:03:10.960 --> 0:03:14.480 stand them. Okay, cool, Yeah, it sounds complex right, So 0:03:14.480 --> 0:03:18.160 speaking of that, how did you get into modeling? So? Um, 0:03:18.240 --> 0:03:21.240 if you can tell by my accent, I'm originally from Australia, 0:03:21.680 --> 0:03:25.760 many many years ago. I was looking for a PhD opportunity, 0:03:26.280 --> 0:03:28.400 a kind of a vague idea of where I wanted 0:03:28.440 --> 0:03:31.960 to go, something energy related, but no, no great drive, 0:03:32.240 --> 0:03:34.320 sort of drifting for a bit, and then applied to 0:03:34.360 --> 0:03:36.440 a bunch of places and bam, I got this offer 0:03:36.480 --> 0:03:39.680 to do a PhD at Oxford. So I couldn't say no. 0:03:40.040 --> 0:03:42.160 It sounds like a pretty sweet place for to end 0:03:42.240 --> 0:03:46.560 up if you're just drifting. I mean, maybe maybe I'm 0:03:46.600 --> 0:03:49.080 under selling myself, but it wasn't. You don't you don't 0:03:49.120 --> 0:03:53.280 just drift into there like that's there's a whole application process, 0:03:53.280 --> 0:03:56.440 like you can't just rock up and anyway. Okay, So 0:03:56.600 --> 0:03:59.720 you start this PC program at Oxford. Yeah, and the 0:03:59.800 --> 0:04:02.960 top pick which I got into was was solar power. 0:04:03.000 --> 0:04:06.120 And before I before I left Australia, I'm joking with 0:04:06.120 --> 0:04:08.839 my supervisors. I'm like, guys, like, I'm not bringing any 0:04:08.880 --> 0:04:10.640 of the sun with me. You realize you have to 0:04:10.640 --> 0:04:13.160 provide that. It actually was. It was ended up being 0:04:13.240 --> 0:04:16.280 quite a significant part of my PhD because we did 0:04:16.360 --> 0:04:19.360 we didn't have sunlight to test a lot of our equipment, 0:04:19.480 --> 0:04:22.520 so I created computer models which could do that for us. 0:04:23.040 --> 0:04:25.640 These are ray tracing models, so you sort of you 0:04:25.680 --> 0:04:28.840 take rays of sunlight, shoot them out, bounce them off 0:04:28.880 --> 0:04:33.479 objects they will like reflect or concentrate, etcetera. So that 0:04:33.480 --> 0:04:36.279 that was probably about half my PhD was sort of 0:04:36.279 --> 0:04:38.800 working on that. That's cool, so you looked at how 0:04:38.920 --> 0:04:42.600 the sun bounces off solar equipment. Basically, yeah, I mean 0:04:42.640 --> 0:04:46.039 we we're working on a solar powered oven. So basically 0:04:46.360 --> 0:04:49.919 a couple of mirrors which combined would result in like 0:04:49.960 --> 0:04:53.360 a very highly concentrated point of sunlight. So we could 0:04:53.360 --> 0:04:57.880 get past three degrees celsius easily and you could not 0:04:58.000 --> 0:05:00.240 just cooking bake, but you could fry as well, which 0:05:00.279 --> 0:05:02.600 was one of the big selling points. So now you're 0:05:02.600 --> 0:05:05.120 at BNF and you switch your focus from from looking 0:05:05.160 --> 0:05:09.159 at how the sun raise bounce to modeling the future 0:05:09.279 --> 0:05:11.760 of power systems? Is that right? That's right? I mean 0:05:12.120 --> 0:05:14.760 one one naturally leads to the other. I assume can 0:05:14.800 --> 0:05:17.840 you talk a bit about that. So you're modeling the future, 0:05:17.880 --> 0:05:20.480 You're you're looking at different scenarios, you're looking at forecasts, 0:05:20.520 --> 0:05:23.360 you're looking at all kinds of things. You could model anything, right, 0:05:23.440 --> 0:05:26.520 why is the power system a good or bad candidate 0:05:26.600 --> 0:05:29.360 for looking at the future. A way to think about 0:05:29.400 --> 0:05:32.680 this is the sort of how in the modeling. So 0:05:32.760 --> 0:05:36.279 there's there's different types of models out there, and some 0:05:36.680 --> 0:05:40.200 are more complicated than others. The simple ones will often 0:05:40.279 --> 0:05:44.240 sort of take an existing data series and just extrapolate 0:05:44.400 --> 0:05:47.200 forward a bit, and they might do some fancy stuff 0:05:47.200 --> 0:05:49.479 with the data to make sure that that that curve 0:05:49.560 --> 0:05:52.359 is accurate, but that that's sort of like a top 0:05:52.400 --> 0:05:56.120 down approach. They're not understanding the fundamental question. The other 0:05:56.160 --> 0:05:58.960 type of model, and the one I work on is 0:05:59.080 --> 0:06:01.440 part of this category, is sort of what we call 0:06:01.520 --> 0:06:06.039 bottom up. So what we interested in is simulating the 0:06:06.120 --> 0:06:08.960 question we're looking at, which for us is the entire 0:06:08.960 --> 0:06:12.120 power system in fact, so we're not we're not blindly 0:06:12.120 --> 0:06:17.599 extrapolating anything. We were creating an entirely new power system, 0:06:17.640 --> 0:06:20.960 a virtual power system in our computer, and and we're 0:06:21.040 --> 0:06:23.520 using that in the model to look at the future. 0:06:23.760 --> 0:06:27.440 All right, So before we started recording here, you told 0:06:27.440 --> 0:06:30.640 me of this super nerdy analogy that that I want 0:06:30.640 --> 0:06:35.480 you to make here about Doctor Strange, the Marvel character. 0:06:35.640 --> 0:06:38.760 So just for the record, everybody, UM, right now, we're London, 0:06:38.800 --> 0:06:41.239 we're on our third lockdown, and my wife has decided 0:06:41.279 --> 0:06:42.960 that we are actually going to go through and watch 0:06:42.960 --> 0:06:45.560 all the Marvel movies right now. She's listening to a 0:06:45.600 --> 0:06:48.480 podcast that goes into great detail on each and we're 0:06:48.520 --> 0:06:50.440 picking him off one by one. She's never seen any 0:06:50.480 --> 0:06:53.719 of them. She's never seen Doctor Strange. So Ian tell 0:06:53.800 --> 0:06:56.840 us about Doctor Strange and his modeling so that there 0:06:56.839 --> 0:06:59.479 will be a spoiler alert here for your wife. Man, 0:07:00.120 --> 0:07:04.360 she didn't listen to the same rate. So in in 0:07:04.360 --> 0:07:07.839 Avengers Endgame, there's this big bad guy. The Avengers are 0:07:07.839 --> 0:07:10.680 trying to beat him. The odds are stacked against them, 0:07:10.720 --> 0:07:13.239 but they've got Doctor Strange on their side, and Doctor 0:07:13.280 --> 0:07:17.080 Strange has this ability to go backwards and forwards in time, 0:07:17.200 --> 0:07:21.120 but also to look at different alternate realities. And he 0:07:21.160 --> 0:07:23.440 goes away, he visits all these alternate realities, and he 0:07:23.480 --> 0:07:26.040 comes back with this sort of really foreboding message that 0:07:26.080 --> 0:07:31.280 I visited fourteen million sive alternate realities. We only beat 0:07:31.480 --> 0:07:33.800 the bad guy. We only beat Sanos in one of them. 0:07:34.080 --> 0:07:38.480 So super long ards, super long ards. But this is 0:07:38.480 --> 0:07:42.120 a really good example of what our model is doing, 0:07:42.960 --> 0:07:46.040 what Nephem is doing sort of behind the scene. So 0:07:47.080 --> 0:07:48.880 you could imagine that you have a model which is 0:07:48.920 --> 0:07:51.520 just mysterious black box. You feed a data gives you 0:07:51.560 --> 0:07:54.240 an output. That's not what we're doing. We're doing we're 0:07:54.240 --> 0:07:57.760 taking the Doctor Strange approach here. So when when we're modeling, 0:07:57.880 --> 0:08:01.560 we're actually creating these sort of alternate realities. We're creating 0:08:01.720 --> 0:08:05.320 different versions of our power system. Um, they will be 0:08:05.360 --> 0:08:07.720 slightly different, they could be very different, but they're all 0:08:07.720 --> 0:08:10.480 different from each other. And we're doing this because we're 0:08:10.520 --> 0:08:13.960 trying to find the version of our power system which 0:08:14.000 --> 0:08:16.120 is the cheapest. The only way we can do that 0:08:16.280 --> 0:08:19.360 is by actually creating them, looking through them, and then 0:08:19.360 --> 0:08:22.840 trying to find the best one. So Doctor Strange had 0:08:22.880 --> 0:08:25.560 about fourteen million. I did some back of the envelope 0:08:25.560 --> 0:08:29.320 calculations before, and I think we did about semi million 0:08:30.080 --> 0:08:36.640 um of these. Yeah, so those are rookie numbers. Doctor Strange. 0:08:37.400 --> 0:08:40.320 Bumpers up. We're crushing Doctor Strange. That's that's awesome. That's 0:08:40.320 --> 0:08:42.800 all we need to know. So we're done, not even 0:08:42.880 --> 0:08:45.240 in the same league. Yeah, that's really cool. So so 0:08:45.280 --> 0:08:49.200 if I get this straight, you are using you're putting 0:08:49.240 --> 0:08:52.200 a bunch of data in and creating alternate realities, you know, 0:08:52.640 --> 0:08:56.280 for the future power system, and then picking the cheapest 0:08:56.320 --> 0:09:00.840 ones to say those could be most likely. Is that yeah? Well, 0:09:00.880 --> 0:09:02.920 I mean we don't make a judgment call about which 0:09:02.960 --> 0:09:06.960 is most likely. That's yeah, that's that's the policy lessons 0:09:07.000 --> 0:09:10.040 that you draw from from the model. But we do 0:09:10.120 --> 0:09:14.679 identify which which is the cheapest Okay, the cheapest to build? Right, Okay? 0:09:14.880 --> 0:09:18.080 When you say identified, like, that's that's the crux of this, 0:09:18.200 --> 0:09:20.840 because even even with all the interns in the world, 0:09:21.240 --> 0:09:24.440 we can't easily manually look through seventy million different power 0:09:24.440 --> 0:09:26.560 systems to find the best one. Yes, so how do 0:09:26.600 --> 0:09:29.800 you do that? So we we've got a solver, which 0:09:29.960 --> 0:09:32.160 which we use in our model what's a solver. So 0:09:32.200 --> 0:09:35.520 a solver is I guess the best way to describe 0:09:35.559 --> 0:09:42.679 it is it's an optimization engine. And generally think about optimization. 0:09:43.400 --> 0:09:47.560 We're normally trying to either minimize or maximize something. For us, 0:09:47.559 --> 0:09:50.360 we're trying to minimize system cost in the power system. 0:09:50.520 --> 0:09:55.720 And so what a solver does is looks at previous data. 0:09:55.920 --> 0:09:58.200 So we look at we'll have a starting set of 0:09:58.240 --> 0:10:02.640 solutions and we'll look at those and then from the 0:10:02.720 --> 0:10:07.400 system costs its seasoned those from the inputs it knows 0:10:07.440 --> 0:10:11.640 it gave them, it can infer what a better guess 0:10:11.800 --> 0:10:15.000 will be the next time around. We do this, okay, 0:10:15.040 --> 0:10:17.520 so it just gets better and better each time. So 0:10:17.600 --> 0:10:20.040 it tweaks something and and says this is higher or 0:10:20.040 --> 0:10:22.320 lower than the best or or how is that right? Yeah? 0:10:22.320 --> 0:10:24.160 And I mean to put it in terms that makes 0:10:24.200 --> 0:10:27.160 sense for for what we're doing, Like the solver might 0:10:27.240 --> 0:10:29.640 try a solution which has a bit more wind and 0:10:29.679 --> 0:10:34.080 a bit more solo. And if you're a faithful subscribe 0:10:34.160 --> 0:10:36.240 a b NF, you you might realize that, hey, those 0:10:36.240 --> 0:10:39.160 technologies are pretty cheap. And if we added that to 0:10:39.200 --> 0:10:42.440 the system and the system costs came down, then the 0:10:42.480 --> 0:10:44.600 solver is going to recognize that, and it's going on, Hey, 0:10:44.720 --> 0:10:47.520 that that that wind and solar solution was pretty good. 0:10:48.240 --> 0:10:50.600 Maybe I'll try some more that is similar to that 0:10:50.760 --> 0:10:52.840 and see if I can't get an even better solution. 0:10:53.160 --> 0:10:55.240 And I imagine there's cases where will go too far 0:10:55.320 --> 0:10:57.560 and say that that too far and made the more 0:10:57.559 --> 0:11:01.000 expensive again, and they'll go back exactly. So we might 0:11:01.000 --> 0:11:04.079 start adding too much, and then we'll start getting curtailment. 0:11:04.200 --> 0:11:07.000 Like some curtailment's fine, but you get to a limit 0:11:07.040 --> 0:11:08.959 where it's just too much, it doesn't make sense, and 0:11:09.040 --> 0:11:11.920 the model just backs off. What's curtailment in case anybody 0:11:11.960 --> 0:11:15.600 isn't now, So, curtailment is when the energy system producing 0:11:15.760 --> 0:11:18.360 more energy than we actually need. So this tends to 0:11:18.440 --> 0:11:20.880 be a problem when we've got things we can't easily 0:11:20.920 --> 0:11:23.920 turn off, like wind or solar. Before we started recording, 0:11:23.960 --> 0:11:26.880 you also talked about an analogy about ants in how 0:11:27.000 --> 0:11:30.000 you build these scenarios. Really, can you tell us a 0:11:30.040 --> 0:11:32.880 bit more about that? Well, this is this is better 0:11:33.440 --> 0:11:36.280 than analogy, even better than an awkward analogy, which is 0:11:36.320 --> 0:11:38.720 my favorite type of analogy. This is this is more 0:11:38.800 --> 0:11:41.720 or less actually what's happening. So so under the hood 0:11:41.760 --> 0:11:45.199 of this solver, the solver is actually based off ant 0:11:45.320 --> 0:11:48.840 behavior in the real world, so it's an ant colony optimizer. 0:11:49.080 --> 0:11:52.960 Is the sort of type of solver the way using 0:11:53.320 --> 0:11:54.880 is that the name we've given it, or is that 0:11:54.920 --> 0:11:56.880 the name that is out in the in the wild, 0:11:57.000 --> 0:11:59.120 in the industry, in the wild. So I mean this 0:11:59.360 --> 0:12:01.800 is this is not our software. This is like commercial 0:12:01.840 --> 0:12:03.640 software that we've bought. But it's very good. It was 0:12:03.640 --> 0:12:06.240 I think it was developed in partnership with the European 0:12:06.320 --> 0:12:11.040 Space Agency and it's holdsome world records for optimal orbital 0:12:11.040 --> 0:12:14.640 flight paths for different Shuttle launches, and I don't I 0:12:14.640 --> 0:12:18.320 don't know exactly. I mean, it's it's it's good stuff though, Okay, amazing. 0:12:18.480 --> 0:12:22.439 I'm convinced. Yeah. The way the way to think of 0:12:22.480 --> 0:12:24.640 what what is happening? If we've got these these these 0:12:24.679 --> 0:12:28.200 alternate realities. The thing is we don't have all seventy 0:12:28.240 --> 0:12:31.080 million of them at once, like we have to fit 0:12:31.120 --> 0:12:35.760 these inside our computers. So normally we're doing, depending on 0:12:35.800 --> 0:12:40.720 the computer, somewhere between four to eighty of these simultaneously 0:12:41.240 --> 0:12:45.360 and what the answer doing, and they're carrying information about 0:12:45.520 --> 0:12:49.840 these alternate realities to and from the solver. So we'll 0:12:49.880 --> 0:12:53.360 create a generation of ants. They'll go out, each ant, 0:12:53.360 --> 0:12:56.839 we'll go to a different reality, report back the system cost, 0:12:57.120 --> 0:13:00.480 and then our solver will create a new generation ants 0:13:01.120 --> 0:13:04.040 based on the information that the previous one provided. And 0:13:04.080 --> 0:13:07.120 at a very high level that that's essentially what's happening. 0:13:07.559 --> 0:13:11.840 And it's the mathematics the underpinn all this are based 0:13:11.880 --> 0:13:15.320 on like path optimization and and the way the ants 0:13:15.320 --> 0:13:17.560 will do that in the real world. And in the 0:13:17.720 --> 0:13:21.000 end you find the least cost futures is that right? 0:13:21.160 --> 0:13:23.120 There will be there will be one lucky ant that 0:13:23.200 --> 0:13:28.120 finds it, one lucky ant, and so let's get into that. 0:13:28.200 --> 0:13:31.600 So you you've found that in this year's edition of 0:13:31.880 --> 0:13:35.960 well Napham the New Energy Forecast Model, but which was 0:13:36.000 --> 0:13:40.200 the basis for the New Energy Outlook, the or NEO, 0:13:41.120 --> 0:13:44.320 which is bens Flagskap report. Is that right? That's right? Okay? 0:13:44.360 --> 0:13:47.439 How does an ant, an individual aunt know when it's 0:13:47.440 --> 0:13:50.320 found the cheapest system cost? So that's a that's a 0:13:50.320 --> 0:13:52.960 good question, because an ant doesn't know how to calculate 0:13:53.000 --> 0:13:56.240 a system cost. So for all the good work that 0:13:56.280 --> 0:13:59.520 the soul is doing, that's that's really just the top 0:13:59.600 --> 0:14:02.920 layer of what's going on in this model. So most 0:14:02.920 --> 0:14:08.840 of the code what's happening is the actual calculation of 0:14:08.960 --> 0:14:13.040 the system cost, and that fundamentally is what NEPHEM is 0:14:13.080 --> 0:14:18.160 the New Energy Forecasting model. It's creating from a given 0:14:18.160 --> 0:14:23.800 set of input data, simulation of the entire power system 0:14:23.840 --> 0:14:26.360 so that we can actually perform that calculation. In the 0:14:26.400 --> 0:14:29.160 methodology document, it says the model solved for a capacity 0:14:29.200 --> 0:14:32.240 mix that minimizes system cost while ensuring I really demand 0:14:32.280 --> 0:14:34.440 is met for an entire year. You Mike, sure all 0:14:34.440 --> 0:14:37.320 the demand is met, putting all that data in, and 0:14:37.360 --> 0:14:40.320 then it says, okay, this is the this is the 0:14:40.440 --> 0:14:44.200 mix that minimizes system costs. Is that right? Yeah? So 0:14:44.640 --> 0:14:46.720 I think maybe the best way to think about it 0:14:46.800 --> 0:14:51.000 is is each and is a different capacity mix. And 0:14:51.040 --> 0:14:54.240 what I mean by capacity mixes we look at all 0:14:54.280 --> 0:14:59.240 the different technologies we have available to us. When solar, coal, gas, nuclear, 0:14:59.720 --> 0:15:04.320 that exactly exactly, and we'll have a different mixture of 0:15:04.360 --> 0:15:07.160 them per ant, So some will have more wind, some 0:15:07.240 --> 0:15:09.520 will have more call, but most of the ants are 0:15:10.000 --> 0:15:12.640 different from each other. So what's happening with this AUNT 0:15:12.760 --> 0:15:16.360 carries the mix, and then Nephem is saying, from this 0:15:16.440 --> 0:15:19.760 given mix of technologies, this is how much the power 0:15:19.800 --> 0:15:23.360 system costs. That's cool, and so then one lucky ant 0:15:23.560 --> 0:15:26.360 is the winner. Well, yeah, I said one lucky ant. 0:15:26.440 --> 0:15:30.160 There's there's a few ants because although those are global analysis, 0:15:30.200 --> 0:15:34.320 we break things down regionally. We don't model every country 0:15:34.360 --> 0:15:37.040 in the world simultaneously, so there's there's quite a few 0:15:37.080 --> 0:15:39.040 ants that win. What regions do you do? How do 0:15:39.040 --> 0:15:43.640 you put it out? So Europe is nine regions, So 0:15:44.120 --> 0:15:49.600 the larger countries we all model individually, so UK, France, Italy, Germany, 0:15:49.760 --> 0:15:53.760 and then the smaller countries we lump together and they'll 0:15:53.960 --> 0:15:56.160 there's a north, south, east, West, and then we do 0:15:56.320 --> 0:15:59.360 iber Area, which is Spain and Portugal together. The US, 0:15:59.520 --> 0:16:03.880 there's thirteen different regions we do that by ISO. China 0:16:04.080 --> 0:16:07.600 is six different regions. India is actually just one region, 0:16:07.720 --> 0:16:11.360 which is our largest region ends up being our largest region. Okay, 0:16:11.360 --> 0:16:13.120 we're going to take a short break and when we 0:16:13.160 --> 0:16:15.560 come back, the new Energy Forecast model with the environment, 0:16:19.040 --> 0:16:20.680 can you tell us a bit more about the data 0:16:20.720 --> 0:16:24.160 that you need in order to run the model and 0:16:24.200 --> 0:16:27.680 how you got it. Yeah, So there's a huge amount 0:16:27.680 --> 0:16:31.040 of data that goes in into this process. And I 0:16:31.120 --> 0:16:33.880 guess the question you're asking is how many data points 0:16:33.880 --> 0:16:38.360 do I need to like synthesize a power system for 0:16:38.400 --> 0:16:40.760 a given region. That's one question, But I'm also just 0:16:40.800 --> 0:16:43.360 really curious about how you went and guard the data, Like, 0:16:44.840 --> 0:16:48.520 I mean, some some of it will come from the terminal, 0:16:48.600 --> 0:16:50.480 some of it's our own data. A lot of the 0:16:50.720 --> 0:16:54.280 cost data, particularly all the renewable cost data, is all internal. 0:16:54.320 --> 0:16:57.160 In fact, I think every cost cost data point is internal. 0:16:57.480 --> 0:17:01.400 There will be commodity prices which come from other teams 0:17:01.640 --> 0:17:05.560 inside ben f and Bloomberg. There will be demand data 0:17:05.720 --> 0:17:10.280 which will come from Bloomberg Economics. There's also demand data 0:17:10.359 --> 0:17:12.800 which comes from the real world, So we'll feed in 0:17:13.040 --> 0:17:17.600 real world demand series for electricity demand from the countries 0:17:17.640 --> 0:17:19.879 where we have that data. How do you go about 0:17:20.160 --> 0:17:23.480 assimilating all this data and planning out the operation of 0:17:23.480 --> 0:17:25.960 the model. So, like, to me a non modeler, that 0:17:26.080 --> 0:17:29.720 just sounds this sounds like a nightmare logistical task or 0:17:29.760 --> 0:17:33.240 planning task. How do you go about planning this model 0:17:33.440 --> 0:17:34.920 for someone who has to do it? It is also 0:17:34.960 --> 0:17:38.879 a nightmare logistical task? Okay, I mean we we so 0:17:38.920 --> 0:17:43.719 we have a database. There's a separate like well defined 0:17:43.800 --> 0:17:47.920 data BRASE which when everything's filled out, that contains every 0:17:47.920 --> 0:17:50.040 single piece of data you need to run the model 0:17:50.200 --> 0:17:53.120 and the processes to get the data in there. There's 0:17:53.119 --> 0:17:55.479 actually a couple of different ones. We use a lot 0:17:55.520 --> 0:18:00.000 of Python scripts to shunt data around, scrape data from website. 0:18:00.040 --> 0:18:03.720 It's there will be some manual points like not the 0:18:03.800 --> 0:18:06.800 big series, but some of the smaller ones will will addit. 0:18:06.840 --> 0:18:10.840 Those the cost data. Again, this is this is all 0:18:11.080 --> 0:18:13.120 a huge data sets that are all coming in because 0:18:13.119 --> 0:18:15.399 when when I say something like the cost of solar, 0:18:15.600 --> 0:18:18.280 it's not just one data point. It will differ by 0:18:18.320 --> 0:18:20.840 country and it differs by year. So we need to 0:18:20.840 --> 0:18:22.240 know that what the cost of solar is in a 0:18:22.240 --> 0:18:26.720 bunch of different countries from now until. Okay, that's no 0:18:26.800 --> 0:18:32.720 small task in it itself, right, No, Yeah, seems pretty intense. Okay, 0:18:33.040 --> 0:18:35.719 I was curious, would you say that is the hardest 0:18:35.720 --> 0:18:38.240 part or what is the hardest part of putting this 0:18:38.320 --> 0:18:41.840 particular model together. This has been so many hard parts. 0:18:41.880 --> 0:18:45.919 I think the storage algorithm was quite quite difficult. Obviously, 0:18:46.000 --> 0:18:49.240 when you're modeling an entire power system, there's different types 0:18:49.280 --> 0:18:52.560 of plants, and they behave differently. A wind or solar 0:18:52.640 --> 0:18:55.320 plant is relatively easy to model from the real world. 0:18:55.560 --> 0:18:58.879 We have this data what window solar generation looks like 0:18:58.920 --> 0:19:02.400 across a given representative weather year, and we can put 0:19:02.480 --> 0:19:04.679 different weather years in the model if we want. But 0:19:04.800 --> 0:19:07.719 for something like storage, like a battery, you don't have 0:19:07.760 --> 0:19:11.399 a well defined data series which is what the output 0:19:11.520 --> 0:19:14.159 is meant to look like. For a battery, the outputs 0:19:14.160 --> 0:19:16.680 the function of what the rest of the system is doing, 0:19:16.720 --> 0:19:19.359 because if there's a lot of cheap energy, or even 0:19:19.359 --> 0:19:22.960 excess energy, then the battery wants to charge during those hours, 0:19:23.320 --> 0:19:25.840 and if there's a shortfall and supply or want to 0:19:25.880 --> 0:19:29.399 discharge in those hours. So that's a completely different type 0:19:29.400 --> 0:19:32.240 of power plant and takes a lot of extra modeling 0:19:32.240 --> 0:19:35.240 effort inside the model to actually calculate. What was the 0:19:35.280 --> 0:19:39.440 most surprising result from the exercise. I think there's a 0:19:39.480 --> 0:19:41.959 lot of cool results that come out of this. I 0:19:42.000 --> 0:19:46.280 think some of the more surprising ones are to do 0:19:46.400 --> 0:19:48.840 with some of the weird scenarios we've done. I'll talk 0:19:48.880 --> 0:19:52.400 about a scenario we we did where we had very 0:19:52.480 --> 0:19:55.760 cheap batteries, and I think most of us would assume 0:19:55.800 --> 0:19:59.359 if you have a power system and you add a 0:19:59.400 --> 0:20:01.960 bunch of cheap our ties, that the winners are probably 0:20:02.000 --> 0:20:05.160 more likely to be renewables. And that's I think it's 0:20:05.160 --> 0:20:07.840 an intuitive thing that that we have, but it's it's 0:20:07.840 --> 0:20:10.280 not necessarily correct, and not at least not in the 0:20:10.320 --> 0:20:13.280 strictest sense, because if you think about a battery, what 0:20:13.320 --> 0:20:16.920 batteries are trying to do is to smooth out to 0:20:17.000 --> 0:20:20.320 supply and demand imbalance. If I had a battery to 0:20:20.440 --> 0:20:24.000 this picture, what happens is that battery will see the 0:20:24.080 --> 0:20:27.080 supplied demand imbalance over the course of the day, and 0:20:27.080 --> 0:20:30.320 the battery says, hey, I want to charge at midday 0:20:30.400 --> 0:20:32.840 when there's like a bunch of extra solar, and then 0:20:32.880 --> 0:20:34.640 what I want to do in the evening is discharge 0:20:34.640 --> 0:20:36.639 because that's where all the demand does. And if you 0:20:36.640 --> 0:20:40.360 do that with a battery, you'll end up raising demand 0:20:40.440 --> 0:20:43.040 from the rest of the system at midday and lowering 0:20:43.080 --> 0:20:46.679 demand in the evening. And hey, like that helps the 0:20:46.720 --> 0:20:49.720 solar guy. It helps the battery, but that's not the 0:20:49.840 --> 0:20:52.240 end of the story, because it also helps the gas plant. 0:20:52.480 --> 0:20:55.320 Without the battery, there's going to be a gas plant 0:20:55.320 --> 0:20:58.560 somewhere in California, which is watching demand fall at midday, 0:20:58.880 --> 0:21:00.560 they're like, do we turn off? We aren't turn off. 0:21:00.600 --> 0:21:05.040 There's like thermodynamic properties inherent to that gas plant that 0:21:05.119 --> 0:21:07.920 means it's very difficult to switch off and on instantly. 0:21:07.960 --> 0:21:10.200 They can't really do it at all. So they're they're 0:21:10.240 --> 0:21:11.960 in a lot of pain at mid day. So they're 0:21:12.000 --> 0:21:15.920 turning their plant right down to the like minimal technical 0:21:16.000 --> 0:21:20.560 feasible limits they can, hoping that like demand doesn't go 0:21:20.600 --> 0:21:22.600 any lower and they have to turn off, and then 0:21:22.640 --> 0:21:25.200 they breathe this huge sigh of relief when demand comes 0:21:25.240 --> 0:21:27.680 back in the evening and they can turn their plant 0:21:27.680 --> 0:21:29.480 back up and and make their money. And when I 0:21:29.520 --> 0:21:32.280