1 00:00:00,200 --> 00:00:02,440 Speaker 1: So right now in my desk have got three books. 2 00:00:02,840 --> 00:00:06,200 Speaker 1: The first one is by David Epstein called Range, How 3 00:00:06,280 --> 00:00:11,080 Speaker 1: Generalists Triumph in a specialized world. Fantastic book. Can't recommend 4 00:00:11,080 --> 00:00:13,399 Speaker 1: it enough. The second book, and at three hundred and 5 00:00:13,400 --> 00:00:15,760 Speaker 1: seventy five page is I mean. Book is a copy 6 00:00:15,760 --> 00:00:18,880 Speaker 1: of Banest's new Energy Outlook twenty twenty or NEO, our 7 00:00:18,960 --> 00:00:21,560 Speaker 1: annual long term scenario analysis on the Future of the 8 00:00:21,640 --> 00:00:24,639 Speaker 1: energy economy, just released in November. It's an Outlook all 9 00:00:24,640 --> 00:00:27,120 Speaker 1: the way to twenty fifty, and his Beenest Flagship Report. 10 00:00:27,680 --> 00:00:29,720 Speaker 1: The third book is called Kings of the Yukon, an 11 00:00:29,720 --> 00:00:33,080 Speaker 1: Alaskan River Journey by Adam Weymouth. Here the author records 12 00:00:33,080 --> 00:00:35,360 Speaker 1: his trip down the Yukon by canoe from Teslin Lake 13 00:00:35,400 --> 00:00:37,760 Speaker 1: in Canada all the way through Alaska to where it 14 00:00:37,840 --> 00:00:40,040 Speaker 1: lets out in the Bearing Sea. He's taking the trip 15 00:00:40,080 --> 00:00:42,760 Speaker 1: to see firsthand the run of the Chinook or king salmon, 16 00:00:42,840 --> 00:00:44,360 Speaker 1: and talk to the people that have been marking the 17 00:00:44,400 --> 00:00:46,919 Speaker 1: time of their annual return for generations. But for me, 18 00:00:47,159 --> 00:00:48,920 Speaker 1: the most fascinating parts of the books so far have 19 00:00:48,960 --> 00:00:51,199 Speaker 1: been when he goes into detail about the behaviors of 20 00:00:51,280 --> 00:00:54,320 Speaker 1: the fish, for example, how they know exactly how to 21 00:00:54,320 --> 00:00:56,880 Speaker 1: get to their spawning grounds down to the tiniest tributary 22 00:00:56,920 --> 00:00:58,800 Speaker 1: or creek of the river without ever having been there 23 00:00:58,840 --> 00:01:01,360 Speaker 1: as adults. Well, today on the show, we're gonna keep 24 00:01:01,400 --> 00:01:04,080 Speaker 1: going with the topic of animal behavior. Turns out the 25 00:01:04,160 --> 00:01:06,920 Speaker 1: engine behind the analysis and NEO, the code that does 26 00:01:06,959 --> 00:01:09,760 Speaker 1: the calculations and creates the scenarios in the outlook, was 27 00:01:09,800 --> 00:01:13,679 Speaker 1: built after animal behaviors, specifically those of ants and birds. 28 00:01:14,160 --> 00:01:17,399 Speaker 1: With us, we've got Ian berriman modeling analyst for PNF. 29 00:01:17,440 --> 00:01:18,800 Speaker 1: He will tell us how the model, known as the 30 00:01:18,840 --> 00:01:21,920 Speaker 1: New Energy Forecast Model or NEPHIM works and how he 31 00:01:21,920 --> 00:01:24,240 Speaker 1: took cues from ants and birds and building it. We'll 32 00:01:24,280 --> 00:01:26,600 Speaker 1: also talk quite a bit about Dr Strange from the 33 00:01:26,640 --> 00:01:31,120 Speaker 1: Marvel movies. So Comberbatch fans get set. Okay. Our discussion 34 00:01:31,160 --> 00:01:34,400 Speaker 1: is based on BNS New Energy Outlook. Being a users 35 00:01:34,440 --> 00:01:36,280 Speaker 1: can get this report on BNF dot com, the BENF 36 00:01:36,360 --> 00:01:38,600 Speaker 1: mobile app, and the Bloomberg terminal. As a reminder, BENIF 37 00:01:38,640 --> 00:01:40,760 Speaker 1: to provide investment or strategy advice, and you can hear 38 00:01:40,760 --> 00:01:42,119 Speaker 1: the full dist claimer at the end of the show. 39 00:01:42,200 --> 00:01:44,280 Speaker 1: I'm Mark Taylor, and you're listening to switch on the 40 00:01:44,319 --> 00:01:55,200 Speaker 1: benf podcast Ian Welcome, thanks for having me. So today 41 00:01:55,240 --> 00:01:57,600 Speaker 1: we're going to talk about modeling. I used to be 42 00:01:57,640 --> 00:01:59,720 Speaker 1: an energy analyst, and I'm not a modeler. You know, 43 00:02:00,040 --> 00:02:02,600 Speaker 1: I'll be I'll be very blunt about that one. But 44 00:02:02,840 --> 00:02:05,200 Speaker 1: for everybody that is listening that doesn't even know what 45 00:02:05,240 --> 00:02:07,200 Speaker 1: a model is, can you just take us all the 46 00:02:07,200 --> 00:02:09,240 Speaker 1: way back to the top and tell us what a 47 00:02:09,240 --> 00:02:11,600 Speaker 1: model is and why they're important. It's a good question. 48 00:02:11,639 --> 00:02:15,000 Speaker 1: So I think why models are important is because we 49 00:02:15,040 --> 00:02:18,600 Speaker 1: don't always have time or space to try everything out. 50 00:02:18,919 --> 00:02:22,919 Speaker 1: In reality, we don't have a physical way of testing 51 00:02:22,960 --> 00:02:26,840 Speaker 1: this hypothesis, don't have the materials. When we're talking about 52 00:02:27,000 --> 00:02:29,800 Speaker 1: models that look towards the future, we don't know what 53 00:02:29,840 --> 00:02:33,600 Speaker 1: that is until it happens. So models help us answer 54 00:02:33,720 --> 00:02:37,639 Speaker 1: questions which we wouldn't otherwise be able to answer. Okay, 55 00:02:37,639 --> 00:02:41,280 Speaker 1: So it helps us get through more data and ask 56 00:02:41,360 --> 00:02:43,639 Speaker 1: more stuff of data that we wouldn't otherwise be able 57 00:02:43,680 --> 00:02:46,799 Speaker 1: to do. Exactly Okay, cool. I think it heard like 58 00:02:46,840 --> 00:02:49,560 Speaker 1: could even be considered like like the model that of 59 00:02:49,600 --> 00:02:51,600 Speaker 1: a building that an architect builds before they actually go 60 00:02:51,600 --> 00:02:53,280 Speaker 1: out and build it, they can see what it looks like. 61 00:02:53,639 --> 00:02:55,960 Speaker 1: You're good to see how it might work exactly. And 62 00:02:56,200 --> 00:02:59,240 Speaker 1: related to that, there will be somewhere on that architect's computer, 63 00:02:59,400 --> 00:03:03,320 Speaker 1: there will be computer model which has all the beams 64 00:03:03,360 --> 00:03:05,280 Speaker 1: and all the weight in the building that is expected 65 00:03:05,280 --> 00:03:07,960 Speaker 1: to carry. And there will be another model which simulates 66 00:03:08,000 --> 00:03:10,840 Speaker 1: winds or earthquakes to make sure the building cap with 67 00:03:10,960 --> 00:03:14,480 Speaker 1: stand them. Okay, cool, Yeah, it sounds complex right, So 68 00:03:14,480 --> 00:03:18,160 Speaker 1: speaking of that, how did you get into modeling? So? Um, 69 00:03:18,240 --> 00:03:21,240 Speaker 1: if you can tell by my accent, I'm originally from Australia, 70 00:03:21,680 --> 00:03:25,760 Speaker 1: many many years ago. I was looking for a PhD opportunity, 71 00:03:26,280 --> 00:03:28,400 Speaker 1: a kind of a vague idea of where I wanted 72 00:03:28,440 --> 00:03:31,960 Speaker 1: to go, something energy related, but no, no great drive, 73 00:03:32,240 --> 00:03:34,320 Speaker 1: sort of drifting for a bit, and then applied to 74 00:03:34,360 --> 00:03:36,440 Speaker 1: a bunch of places and bam, I got this offer 75 00:03:36,480 --> 00:03:39,680 Speaker 1: to do a PhD at Oxford. So I couldn't say no. 76 00:03:40,040 --> 00:03:42,160 Speaker 1: It sounds like a pretty sweet place for to end 77 00:03:42,240 --> 00:03:46,560 Speaker 1: up if you're just drifting. I mean, maybe maybe I'm 78 00:03:46,600 --> 00:03:49,080 Speaker 1: under selling myself, but it wasn't. You don't you don't 79 00:03:49,120 --> 00:03:53,280 Speaker 1: just drift into there like that's there's a whole application process, 80 00:03:53,280 --> 00:03:56,440 Speaker 1: like you can't just rock up and anyway. Okay, So 81 00:03:56,600 --> 00:03:59,720 Speaker 1: you start this PC program at Oxford. Yeah, and the 82 00:03:59,800 --> 00:04:02,960 Speaker 1: top pick which I got into was was solar power. 83 00:04:03,000 --> 00:04:06,120 Speaker 1: And before I before I left Australia, I'm joking with 84 00:04:06,120 --> 00:04:08,839 Speaker 1: my supervisors. I'm like, guys, like, I'm not bringing any 85 00:04:08,880 --> 00:04:10,640 Speaker 1: of the sun with me. You realize you have to 86 00:04:10,640 --> 00:04:13,160 Speaker 1: provide that. It actually was. It was ended up being 87 00:04:13,240 --> 00:04:16,280 Speaker 1: quite a significant part of my PhD because we did 88 00:04:16,360 --> 00:04:19,360 Speaker 1: we didn't have sunlight to test a lot of our equipment, 89 00:04:19,480 --> 00:04:22,520 Speaker 1: so I created computer models which could do that for us. 90 00:04:23,040 --> 00:04:25,640 Speaker 1: These are ray tracing models, so you sort of you 91 00:04:25,680 --> 00:04:28,840 Speaker 1: take rays of sunlight, shoot them out, bounce them off 92 00:04:28,880 --> 00:04:33,479 Speaker 1: objects they will like reflect or concentrate, etcetera. So that 93 00:04:33,480 --> 00:04:36,279 Speaker 1: that was probably about half my PhD was sort of 94 00:04:36,279 --> 00:04:38,800 Speaker 1: working on that. That's cool, so you looked at how 95 00:04:38,920 --> 00:04:42,600 Speaker 1: the sun bounces off solar equipment. Basically, yeah, I mean 96 00:04:42,640 --> 00:04:46,039 Speaker 1: we we're working on a solar powered oven. So basically 97 00:04:46,360 --> 00:04:49,919 Speaker 1: a couple of mirrors which combined would result in like 98 00:04:49,960 --> 00:04:53,360 Speaker 1: a very highly concentrated point of sunlight. So we could 99 00:04:53,360 --> 00:04:57,880 Speaker 1: get past three degrees celsius easily and you could not 100 00:04:58,000 --> 00:05:00,240 Speaker 1: just cooking bake, but you could fry as well, which 101 00:05:00,279 --> 00:05:02,600 Speaker 1: was one of the big selling points. So now you're 102 00:05:02,600 --> 00:05:05,120 Speaker 1: at BNF and you switch your focus from from looking 103 00:05:05,160 --> 00:05:09,159 Speaker 1: at how the sun raise bounce to modeling the future 104 00:05:09,279 --> 00:05:11,760 Speaker 1: of power systems? Is that right? That's right? I mean 105 00:05:12,120 --> 00:05:14,760 Speaker 1: one one naturally leads to the other. I assume can 106 00:05:14,800 --> 00:05:17,840 Speaker 1: you talk a bit about that. So you're modeling the future, 107 00:05:17,880 --> 00:05:20,480 Speaker 1: You're you're looking at different scenarios, you're looking at forecasts, 108 00:05:20,520 --> 00:05:23,360 Speaker 1: you're looking at all kinds of things. You could model anything, right, 109 00:05:23,440 --> 00:05:26,520 Speaker 1: why is the power system a good or bad candidate 110 00:05:26,600 --> 00:05:29,360 Speaker 1: for looking at the future. A way to think about 111 00:05:29,400 --> 00:05:32,680 Speaker 1: this is the sort of how in the modeling. So 112 00:05:32,760 --> 00:05:36,279 Speaker 1: there's there's different types of models out there, and some 113 00:05:36,680 --> 00:05:40,200 Speaker 1: are more complicated than others. The simple ones will often 114 00:05:40,279 --> 00:05:44,240 Speaker 1: sort of take an existing data series and just extrapolate 115 00:05:44,400 --> 00:05:47,200 Speaker 1: forward a bit, and they might do some fancy stuff 116 00:05:47,200 --> 00:05:49,479 Speaker 1: with the data to make sure that that that curve 117 00:05:49,560 --> 00:05:52,359 Speaker 1: is accurate, but that that's sort of like a top 118 00:05:52,400 --> 00:05:56,120 Speaker 1: down approach. They're not understanding the fundamental question. The other 119 00:05:56,160 --> 00:05:58,960 Speaker 1: type of model, and the one I work on is 120 00:05:59,080 --> 00:06:01,440 Speaker 1: part of this category, is sort of what we call 121 00:06:01,520 --> 00:06:06,039 Speaker 1: bottom up. So what we interested in is simulating the 122 00:06:06,120 --> 00:06:08,960 Speaker 1: question we're looking at, which for us is the entire 123 00:06:08,960 --> 00:06:12,120 Speaker 1: power system in fact, so we're not we're not blindly 124 00:06:12,120 --> 00:06:17,599 Speaker 1: extrapolating anything. We were creating an entirely new power system, 125 00:06:17,640 --> 00:06:20,960 Speaker 1: a virtual power system in our computer, and and we're 126 00:06:21,040 --> 00:06:23,520 Speaker 1: using that in the model to look at the future. 127 00:06:23,760 --> 00:06:27,440 Speaker 1: All right, So before we started recording here, you told 128 00:06:27,440 --> 00:06:30,640 Speaker 1: me of this super nerdy analogy that that I want 129 00:06:30,640 --> 00:06:35,480 Speaker 1: you to make here about Doctor Strange, the Marvel character. 130 00:06:35,640 --> 00:06:38,760 Speaker 1: So just for the record, everybody, UM, right now, we're London, 131 00:06:38,800 --> 00:06:41,239 Speaker 1: we're on our third lockdown, and my wife has decided 132 00:06:41,279 --> 00:06:42,960 Speaker 1: that we are actually going to go through and watch 133 00:06:42,960 --> 00:06:45,560 Speaker 1: all the Marvel movies right now. She's listening to a 134 00:06:45,600 --> 00:06:48,480 Speaker 1: podcast that goes into great detail on each and we're 135 00:06:48,520 --> 00:06:50,440 Speaker 1: picking him off one by one. She's never seen any 136 00:06:50,480 --> 00:06:53,719 Speaker 1: of them. She's never seen Doctor Strange. So Ian tell 137 00:06:53,800 --> 00:06:56,840 Speaker 1: us about Doctor Strange and his modeling so that there 138 00:06:56,839 --> 00:06:59,479 Speaker 1: will be a spoiler alert here for your wife. Man, 139 00:07:00,120 --> 00:07:04,360 Speaker 1: she didn't listen to the same rate. So in in 140 00:07:04,360 --> 00:07:07,839 Speaker 1: Avengers Endgame, there's this big bad guy. The Avengers are 141 00:07:07,839 --> 00:07:10,680 Speaker 1: trying to beat him. The odds are stacked against them, 142 00:07:10,720 --> 00:07:13,239 Speaker 1: but they've got Doctor Strange on their side, and Doctor 143 00:07:13,280 --> 00:07:17,080 Speaker 1: Strange has this ability to go backwards and forwards in time, 144 00:07:17,200 --> 00:07:21,120 Speaker 1: but also to look at different alternate realities. And he 145 00:07:21,160 --> 00:07:23,440 Speaker 1: goes away, he visits all these alternate realities, and he 146 00:07:23,480 --> 00:07:26,040 Speaker 1: comes back with this sort of really foreboding message that 147 00:07:26,080 --> 00:07:31,280 Speaker 1: I visited fourteen million sive alternate realities. We only beat 148 00:07:31,480 --> 00:07:33,800 Speaker 1: the bad guy. We only beat Sanos in one of them. 149 00:07:34,080 --> 00:07:38,480 Speaker 1: So super long ards, super long ards. But this is 150 00:07:38,480 --> 00:07:42,120 Speaker 1: a really good example of what our model is doing, 151 00:07:42,960 --> 00:07:46,040 Speaker 1: what Nephem is doing sort of behind the scene. So 152 00:07:47,080 --> 00:07:48,880 Speaker 1: you could imagine that you have a model which is 153 00:07:48,920 --> 00:07:51,520 Speaker 1: just mysterious black box. You feed a data gives you 154 00:07:51,560 --> 00:07:54,240 Speaker 1: an output. That's not what we're doing. We're doing we're 155 00:07:54,240 --> 00:07:57,760 Speaker 1: taking the Doctor Strange approach here. So when when we're modeling, 156 00:07:57,880 --> 00:08:01,560 Speaker 1: we're actually creating these sort of alternate realities. We're creating 157 00:08:01,720 --> 00:08:05,320 Speaker 1: different versions of our power system. Um, they will be 158 00:08:05,360 --> 00:08:07,720 Speaker 1: slightly different, they could be very different, but they're all 159 00:08:07,720 --> 00:08:10,480 Speaker 1: different from each other. And we're doing this because we're 160 00:08:10,520 --> 00:08:13,960 Speaker 1: trying to find the version of our power system which 161 00:08:14,000 --> 00:08:16,120 Speaker 1: is the cheapest. The only way we can do that 162 00:08:16,280 --> 00:08:19,360 Speaker 1: is by actually creating them, looking through them, and then 163 00:08:19,360 --> 00:08:22,840 Speaker 1: trying to find the best one. So Doctor Strange had 164 00:08:22,880 --> 00:08:25,560 Speaker 1: about fourteen million. I did some back of the envelope 165 00:08:25,560 --> 00:08:29,320 Speaker 1: calculations before, and I think we did about semi million 166 00:08:30,080 --> 00:08:36,640 Speaker 1: um of these. Yeah, so those are rookie numbers. Doctor Strange. 167 00:08:37,400 --> 00:08:40,320 Speaker 1: Bumpers up. We're crushing Doctor Strange. That's that's awesome. That's 168 00:08:40,320 --> 00:08:42,800 Speaker 1: all we need to know. So we're done, not even 169 00:08:42,880 --> 00:08:45,240 Speaker 1: in the same league. Yeah, that's really cool. So so 170 00:08:45,280 --> 00:08:49,200 Speaker 1: if I get this straight, you are using you're putting 171 00:08:49,240 --> 00:08:52,200 Speaker 1: a bunch of data in and creating alternate realities, you know, 172 00:08:52,640 --> 00:08:56,280 Speaker 1: for the future power system, and then picking the cheapest 173 00:08:56,320 --> 00:09:00,840 Speaker 1: ones to say those could be most likely. Is that yeah? Well, 174 00:09:00,880 --> 00:09:02,920 Speaker 1: I mean we don't make a judgment call about which 175 00:09:02,960 --> 00:09:06,960 Speaker 1: is most likely. That's yeah, that's that's the policy lessons 176 00:09:07,000 --> 00:09:10,040 Speaker 1: that you draw from from the model. But we do 177 00:09:10,120 --> 00:09:14,679 Speaker 1: identify which which is the cheapest Okay, the cheapest to build? Right, Okay? 178 00:09:14,880 --> 00:09:18,080 Speaker 1: When you say identified, like, that's that's the crux of this, 179 00:09:18,200 --> 00:09:20,840 Speaker 1: because even even with all the interns in the world, 180 00:09:21,240 --> 00:09:24,440 Speaker 1: we can't easily manually look through seventy million different power 181 00:09:24,440 --> 00:09:26,560 Speaker 1: systems to find the best one. Yes, so how do 182 00:09:26,600 --> 00:09:29,800 Speaker 1: you do that? So we we've got a solver, which 183 00:09:29,960 --> 00:09:32,160 Speaker 1: which we use in our model what's a solver. So 184 00:09:32,200 --> 00:09:35,520 Speaker 1: a solver is I guess the best way to describe 185 00:09:35,559 --> 00:09:42,679 Speaker 1: it is it's an optimization engine. And generally think about optimization. 186 00:09:43,400 --> 00:09:47,560 Speaker 1: We're normally trying to either minimize or maximize something. For us, 187 00:09:47,559 --> 00:09:50,360 Speaker 1: we're trying to minimize system cost in the power system. 188 00:09:50,520 --> 00:09:55,720 Speaker 1: And so what a solver does is looks at previous data. 189 00:09:55,920 --> 00:09:58,200 Speaker 1: So we look at we'll have a starting set of 190 00:09:58,240 --> 00:10:02,640 Speaker 1: solutions and we'll look at those and then from the 191 00:10:02,720 --> 00:10:07,400 Speaker 1: system costs its seasoned those from the inputs it knows 192 00:10:07,440 --> 00:10:11,640 Speaker 1: it gave them, it can infer what a better guess 193 00:10:11,800 --> 00:10:15,000 Speaker 1: will be the next time around. We do this, okay, 194 00:10:15,040 --> 00:10:17,520 Speaker 1: so it just gets better and better each time. So 195 00:10:17,600 --> 00:10:20,040 Speaker 1: it tweaks something and and says this is higher or 196 00:10:20,040 --> 00:10:22,320 Speaker 1: lower than the best or or how is that right? Yeah? 197 00:10:22,320 --> 00:10:24,160 Speaker 1: And I mean to put it in terms that makes 198 00:10:24,200 --> 00:10:27,160 Speaker 1: sense for for what we're doing, Like the solver might 199 00:10:27,240 --> 00:10:29,640 Speaker 1: try a solution which has a bit more wind and 200 00:10:29,679 --> 00:10:34,080 Speaker 1: a bit more solo. And if you're a faithful subscribe 201 00:10:34,160 --> 00:10:36,240 Speaker 1: a b NF, you you might realize that, hey, those 202 00:10:36,240 --> 00:10:39,160 Speaker 1: technologies are pretty cheap. And if we added that to 203 00:10:39,200 --> 00:10:42,440 Speaker 1: the system and the system costs came down, then the 204 00:10:42,480 --> 00:10:44,600 Speaker 1: solver is going to recognize that, and it's going on, Hey, 205 00:10:44,720 --> 00:10:47,520 Speaker 1: that that that wind and solar solution was pretty good. 206 00:10:48,240 --> 00:10:50,600 Speaker 1: Maybe I'll try some more that is similar to that 207 00:10:50,760 --> 00:10:52,840 Speaker 1: and see if I can't get an even better solution. 208 00:10:53,160 --> 00:10:55,240 Speaker 1: And I imagine there's cases where will go too far 209 00:10:55,320 --> 00:10:57,560 Speaker 1: and say that that too far and made the more 210 00:10:57,559 --> 00:11:01,000 Speaker 1: expensive again, and they'll go back exactly. So we might 211 00:11:01,000 --> 00:11:04,079 Speaker 1: start adding too much, and then we'll start getting curtailment. 212 00:11:04,200 --> 00:11:07,000 Speaker 1: Like some curtailment's fine, but you get to a limit 213 00:11:07,040 --> 00:11:08,959 Speaker 1: where it's just too much, it doesn't make sense, and 214 00:11:09,040 --> 00:11:11,920 Speaker 1: the model just backs off. What's curtailment in case anybody 215 00:11:11,960 --> 00:11:15,600 Speaker 1: isn't now, So, curtailment is when the energy system producing 216 00:11:15,760 --> 00:11:18,360 Speaker 1: more energy than we actually need. So this tends to 217 00:11:18,440 --> 00:11:20,880 Speaker 1: be a problem when we've got things we can't easily 218 00:11:20,920 --> 00:11:23,920 Speaker 1: turn off, like wind or solar. Before we started recording, 219 00:11:23,960 --> 00:11:26,880 Speaker 1: you also talked about an analogy about ants in how 220 00:11:27,000 --> 00:11:30,000 Speaker 1: you build these scenarios. Really, can you tell us a 221 00:11:30,040 --> 00:11:32,880 Speaker 1: bit more about that? Well, this is this is better 222 00:11:33,440 --> 00:11:36,280 Speaker 1: than analogy, even better than an awkward analogy, which is 223 00:11:36,320 --> 00:11:38,720 Speaker 1: my favorite type of analogy. This is this is more 224 00:11:38,800 --> 00:11:41,720 Speaker 1: or less actually what's happening. So so under the hood 225 00:11:41,760 --> 00:11:45,199 Speaker 1: of this solver, the solver is actually based off ant 226 00:11:45,320 --> 00:11:48,840 Speaker 1: behavior in the real world, so it's an ant colony optimizer. 227 00:11:49,080 --> 00:11:52,960 Speaker 1: Is the sort of type of solver the way using 228 00:11:53,320 --> 00:11:54,880 Speaker 1: is that the name we've given it, or is that 229 00:11:54,920 --> 00:11:56,880 Speaker 1: the name that is out in the in the wild, 230 00:11:57,000 --> 00:11:59,120 Speaker 1: in the industry, in the wild. So I mean this 231 00:11:59,360 --> 00:12:01,800 Speaker 1: is this is not our software. This is like commercial 232 00:12:01,840 --> 00:12:03,640 Speaker 1: software that we've bought. But it's very good. It was 233 00:12:03,640 --> 00:12:06,240 Speaker 1: I think it was developed in partnership with the European 234 00:12:06,320 --> 00:12:11,040 Speaker 1: Space Agency and it's holdsome world records for optimal orbital 235 00:12:11,040 --> 00:12:14,640 Speaker 1: flight paths for different Shuttle launches, and I don't I 236 00:12:14,640 --> 00:12:18,320 Speaker 1: don't know exactly. I mean, it's it's it's good stuff though, Okay, amazing. 237 00:12:18,480 --> 00:12:22,439 Speaker 1: I'm convinced. Yeah. The way the way to think of 238 00:12:22,480 --> 00:12:24,640 Speaker 1: what what is happening? If we've got these these these 239 00:12:24,679 --> 00:12:28,200 Speaker 1: alternate realities. The thing is we don't have all seventy 240 00:12:28,240 --> 00:12:31,080 Speaker 1: million of them at once, like we have to fit 241 00:12:31,120 --> 00:12:35,760 Speaker 1: these inside our computers. So normally we're doing, depending on 242 00:12:35,800 --> 00:12:40,720 Speaker 1: the computer, somewhere between four to eighty of these simultaneously 243 00:12:41,240 --> 00:12:45,360 Speaker 1: and what the answer doing, and they're carrying information about 244 00:12:45,520 --> 00:12:49,840 Speaker 1: these alternate realities to and from the solver. So we'll 245 00:12:49,880 --> 00:12:53,360 Speaker 1: create a generation of ants. They'll go out, each ant, 246 00:12:53,360 --> 00:12:56,839 Speaker 1: we'll go to a different reality, report back the system cost, 247 00:12:57,120 --> 00:13:00,480 Speaker 1: and then our solver will create a new generation ants 248 00:13:01,120 --> 00:13:04,040 Speaker 1: based on the information that the previous one provided. And 249 00:13:04,080 --> 00:13:07,120 Speaker 1: at a very high level that that's essentially what's happening. 250 00:13:07,559 --> 00:13:11,840 Speaker 1: And it's the mathematics the underpinn all this are based 251 00:13:11,880 --> 00:13:15,320 Speaker 1: on like path optimization and and the way the ants 252 00:13:15,320 --> 00:13:17,560 Speaker 1: will do that in the real world. And in the 253 00:13:17,720 --> 00:13:21,000 Speaker 1: end you find the least cost futures is that right? 254 00:13:21,160 --> 00:13:23,120 Speaker 1: There will be there will be one lucky ant that 255 00:13:23,200 --> 00:13:28,120 Speaker 1: finds it, one lucky ant, and so let's get into that. 256 00:13:28,200 --> 00:13:31,600 Speaker 1: So you you've found that in this year's edition of 257 00:13:31,880 --> 00:13:35,960 Speaker 1: well Napham the New Energy Forecast Model, but which was 258 00:13:36,000 --> 00:13:40,200 Speaker 1: the basis for the New Energy Outlook, the or NEO, 259 00:13:41,120 --> 00:13:44,320 Speaker 1: which is bens Flagskap report. Is that right? That's right? Okay? 260 00:13:44,360 --> 00:13:47,439 Speaker 1: How does an ant, an individual aunt know when it's 261 00:13:47,440 --> 00:13:50,320 Speaker 1: found the cheapest system cost? So that's a that's a 262 00:13:50,320 --> 00:13:52,960 Speaker 1: good question, because an ant doesn't know how to calculate 263 00:13:53,000 --> 00:13:56,240 Speaker 1: a system cost. So for all the good work that 264 00:13:56,280 --> 00:13:59,520 Speaker 1: the soul is doing, that's that's really just the top 265 00:13:59,600 --> 00:14:02,920 Speaker 1: layer of what's going on in this model. So most 266 00:14:02,920 --> 00:14:08,840 Speaker 1: of the code what's happening is the actual calculation of 267 00:14:08,960 --> 00:14:13,040 Speaker 1: the system cost, and that fundamentally is what NEPHEM is 268 00:14:13,080 --> 00:14:18,160 Speaker 1: the New Energy Forecasting model. It's creating from a given 269 00:14:18,160 --> 00:14:23,800 Speaker 1: set of input data, simulation of the entire power system 270 00:14:23,840 --> 00:14:26,360 Speaker 1: so that we can actually perform that calculation. In the 271 00:14:26,400 --> 00:14:29,160 Speaker 1: methodology document, it says the model solved for a capacity 272 00:14:29,200 --> 00:14:32,240 Speaker 1: mix that minimizes system cost while ensuring I really demand 273 00:14:32,280 --> 00:14:34,440 Speaker 1: is met for an entire year. You Mike, sure all 274 00:14:34,440 --> 00:14:37,320 Speaker 1: the demand is met, putting all that data in, and 275 00:14:37,360 --> 00:14:40,320 Speaker 1: then it says, okay, this is the this is the 276 00:14:40,440 --> 00:14:44,200 Speaker 1: mix that minimizes system costs. Is that right? Yeah? So 277 00:14:44,640 --> 00:14:46,720 Speaker 1: I think maybe the best way to think about it 278 00:14:46,800 --> 00:14:51,000 Speaker 1: is is each and is a different capacity mix. And 279 00:14:51,040 --> 00:14:54,240 Speaker 1: what I mean by capacity mixes we look at all 280 00:14:54,280 --> 00:14:59,240 Speaker 1: the different technologies we have available to us. When solar, coal, gas, nuclear, 281 00:14:59,720 --> 00:15:04,320 Speaker 1: that exactly exactly, and we'll have a different mixture of 282 00:15:04,360 --> 00:15:07,160 Speaker 1: them per ant, So some will have more wind, some 283 00:15:07,240 --> 00:15:09,520 Speaker 1: will have more call, but most of the ants are 284 00:15:10,000 --> 00:15:12,640 Speaker 1: different from each other. So what's happening with this AUNT 285 00:15:12,760 --> 00:15:16,360 Speaker 1: carries the mix, and then Nephem is saying, from this 286 00:15:16,440 --> 00:15:19,760 Speaker 1: given mix of technologies, this is how much the power 287 00:15:19,800 --> 00:15:23,360 Speaker 1: system costs. That's cool, and so then one lucky ant 288 00:15:23,560 --> 00:15:26,360 Speaker 1: is the winner. Well, yeah, I said one lucky ant. 289 00:15:26,440 --> 00:15:30,160 Speaker 1: There's there's a few ants because although those are global analysis, 290 00:15:30,200 --> 00:15:34,320 Speaker 1: we break things down regionally. We don't model every country 291 00:15:34,360 --> 00:15:37,040 Speaker 1: in the world simultaneously, so there's there's quite a few 292 00:15:37,080 --> 00:15:39,040 Speaker 1: ants that win. What regions do you do? How do 293 00:15:39,040 --> 00:15:43,640 Speaker 1: you put it out? So Europe is nine regions, So 294 00:15:44,120 --> 00:15:49,600 Speaker 1: the larger countries we all model individually, so UK, France, Italy, Germany, 295 00:15:49,760 --> 00:15:53,760 Speaker 1: and then the smaller countries we lump together and they'll 296 00:15:53,960 --> 00:15:56,160 Speaker 1: there's a north, south, east, West, and then we do 297 00:15:56,320 --> 00:15:59,360 Speaker 1: iber Area, which is Spain and Portugal together. The US, 298 00:15:59,520 --> 00:16:03,880 Speaker 1: there's thirteen different regions we do that by ISO. China 299 00:16:04,080 --> 00:16:07,600 Speaker 1: is six different regions. India is actually just one region, 300 00:16:07,720 --> 00:16:11,360 Speaker 1: which is our largest region ends up being our largest region. Okay, 301 00:16:11,360 --> 00:16:13,120 Speaker 1: we're going to take a short break and when we 302 00:16:13,160 --> 00:16:15,560 Speaker 1: come back, the new Energy Forecast model with the environment, 303 00:16:19,040 --> 00:16:20,680 Speaker 1: can you tell us a bit more about the data 304 00:16:20,720 --> 00:16:24,160 Speaker 1: that you need in order to run the model and 305 00:16:24,200 --> 00:16:27,680 Speaker 1: how you got it. Yeah, So there's a huge amount 306 00:16:27,680 --> 00:16:31,040 Speaker 1: of data that goes in into this process. And I 307 00:16:31,120 --> 00:16:33,880 Speaker 1: guess the question you're asking is how many data points 308 00:16:33,880 --> 00:16:38,360 Speaker 1: do I need to like synthesize a power system for 309 00:16:38,400 --> 00:16:40,760 Speaker 1: a given region. That's one question, But I'm also just 310 00:16:40,800 --> 00:16:43,360 Speaker 1: really curious about how you went and guard the data, Like, 311 00:16:44,840 --> 00:16:48,520 Speaker 1: I mean, some some of it will come from the terminal, 312 00:16:48,600 --> 00:16:50,480 Speaker 1: some of it's our own data. A lot of the 313 00:16:50,720 --> 00:16:54,280 Speaker 1: cost data, particularly all the renewable cost data, is all internal. 314 00:16:54,320 --> 00:16:57,160 Speaker 1: In fact, I think every cost cost data point is internal. 315 00:16:57,480 --> 00:17:01,400 Speaker 1: There will be commodity prices which come from other teams 316 00:17:01,640 --> 00:17:05,560 Speaker 1: inside ben f and Bloomberg. There will be demand data 317 00:17:05,720 --> 00:17:10,280 Speaker 1: which will come from Bloomberg Economics. There's also demand data 318 00:17:10,359 --> 00:17:12,800 Speaker 1: which comes from the real world, So we'll feed in 319 00:17:13,040 --> 00:17:17,600 Speaker 1: real world demand series for electricity demand from the countries 320 00:17:17,640 --> 00:17:19,879 Speaker 1: where we have that data. How do you go about 321 00:17:20,160 --> 00:17:23,480 Speaker 1: assimilating all this data and planning out the operation of 322 00:17:23,480 --> 00:17:25,960 Speaker 1: the model. So, like, to me a non modeler, that 323 00:17:26,080 --> 00:17:29,720 Speaker 1: just sounds this sounds like a nightmare logistical task or 324 00:17:29,760 --> 00:17:33,240 Speaker 1: planning task. How do you go about planning this model 325 00:17:33,440 --> 00:17:34,920 Speaker 1: for someone who has to do it? It is also 326 00:17:34,960 --> 00:17:38,879 Speaker 1: a nightmare logistical task? Okay, I mean we we so 327 00:17:38,920 --> 00:17:43,719 Speaker 1: we have a database. There's a separate like well defined 328 00:17:43,800 --> 00:17:47,920 Speaker 1: data BRASE which when everything's filled out, that contains every 329 00:17:47,920 --> 00:17:50,040 Speaker 1: single piece of data you need to run the model 330 00:17:50,200 --> 00:17:53,120 Speaker 1: and the processes to get the data in there. There's 331 00:17:53,119 --> 00:17:55,479 Speaker 1: actually a couple of different ones. We use a lot 332 00:17:55,520 --> 00:18:00,000 Speaker 1: of Python scripts to shunt data around, scrape data from website. 333 00:18:00,040 --> 00:18:03,720 Speaker 1: It's there will be some manual points like not the 334 00:18:03,800 --> 00:18:06,800 Speaker 1: big series, but some of the smaller ones will will addit. 335 00:18:06,840 --> 00:18:10,840 Speaker 1: Those the cost data. Again, this is this is all 336 00:18:11,080 --> 00:18:13,120 Speaker 1: a huge data sets that are all coming in because 337 00:18:13,119 --> 00:18:15,399 Speaker 1: when when I say something like the cost of solar, 338 00:18:15,600 --> 00:18:18,280 Speaker 1: it's not just one data point. It will differ by 339 00:18:18,320 --> 00:18:20,840 Speaker 1: country and it differs by year. So we need to 340 00:18:20,840 --> 00:18:22,240 Speaker 1: know that what the cost of solar is in a 341 00:18:22,240 --> 00:18:26,720 Speaker 1: bunch of different countries from now until. Okay, that's no 342 00:18:26,800 --> 00:18:32,720 Speaker 1: small task in it itself, right, No, Yeah, seems pretty intense. Okay, 343 00:18:33,040 --> 00:18:35,719 Speaker 1: I was curious, would you say that is the hardest 344 00:18:35,720 --> 00:18:38,240 Speaker 1: part or what is the hardest part of putting this 345 00:18:38,320 --> 00:18:41,840 Speaker 1: particular model together. This has been so many hard parts. 346 00:18:41,880 --> 00:18:45,919 Speaker 1: I think the storage algorithm was quite quite difficult. Obviously, 347 00:18:46,000 --> 00:18:49,240 Speaker 1: when you're modeling an entire power system, there's different types 348 00:18:49,280 --> 00:18:52,560 Speaker 1: of plants, and they behave differently. A wind or solar 349 00:18:52,640 --> 00:18:55,320 Speaker 1: plant is relatively easy to model from the real world. 350 00:18:55,560 --> 00:18:58,879 Speaker 1: We have this data what window solar generation looks like 351 00:18:58,920 --> 00:19:02,400 Speaker 1: across a given representative weather year, and we can put 352 00:19:02,480 --> 00:19:04,679 Speaker 1: different weather years in the model if we want. But 353 00:19:04,800 --> 00:19:07,719 Speaker 1: for something like storage, like a battery, you don't have 354 00:19:07,760 --> 00:19:11,399 Speaker 1: a well defined data series which is what the output 355 00:19:11,520 --> 00:19:14,159 Speaker 1: is meant to look like. For a battery, the outputs 356 00:19:14,160 --> 00:19:16,680 Speaker 1: the function of what the rest of the system is doing, 357 00:19:16,720 --> 00:19:19,359 Speaker 1: because if there's a lot of cheap energy, or even 358 00:19:19,359 --> 00:19:22,960 Speaker 1: excess energy, then the battery wants to charge during those hours, 359 00:19:23,320 --> 00:19:25,840 Speaker 1: and if there's a shortfall and supply or want to 360 00:19:25,880 --> 00:19:29,399 Speaker 1: discharge in those hours. So that's a completely different type 361 00:19:29,400 --> 00:19:32,240 Speaker 1: of power plant and takes a lot of extra modeling 362 00:19:32,240 --> 00:19:35,240 Speaker 1: effort inside the model to actually calculate. What was the 363 00:19:35,280 --> 00:19:39,440 Speaker 1: most surprising result from the exercise. I think there's a 364 00:19:39,480 --> 00:19:41,959 Speaker 1: lot of cool results that come out of this. I 365 00:19:42,000 --> 00:19:46,280 Speaker 1: think some of the more surprising ones are to do 366 00:19:46,400 --> 00:19:48,840 Speaker 1: with some of the weird scenarios we've done. I'll talk 367 00:19:48,880 --> 00:19:52,400 Speaker 1: about a scenario we we did where we had very 368 00:19:52,480 --> 00:19:55,760 Speaker 1: cheap batteries, and I think most of us would assume 369 00:19:55,800 --> 00:19:59,359 Speaker 1: if you have a power system and you add a 370 00:19:59,400 --> 00:20:01,960 Speaker 1: bunch of cheap our ties, that the winners are probably 371 00:20:02,000 --> 00:20:05,160 Speaker 1: more likely to be renewables. And that's I think it's 372 00:20:05,160 --> 00:20:07,840 Speaker 1: an intuitive thing that that we have, but it's it's 373 00:20:07,840 --> 00:20:10,280 Speaker 1: not necessarily correct, and not at least not in the 374 00:20:10,320 --> 00:20:13,280 Speaker 1: strictest sense, because if you think about a battery, what 375 00:20:13,320 --> 00:20:16,920 Speaker 1: batteries are trying to do is to smooth out to 376 00:20:17,000 --> 00:20:20,320 Speaker 1: supply and demand imbalance. If I had a battery to 377 00:20:20,440 --> 00:20:24,000 Speaker 1: this picture, what happens is that battery will see the 378 00:20:24,080 --> 00:20:27,080 Speaker 1: supplied demand imbalance over the course of the day, and 379 00:20:27,080 --> 00:20:30,320 Speaker 1: the battery says, hey, I want to charge at midday 380 00:20:30,400 --> 00:20:32,840 Speaker 1: when there's like a bunch of extra solar, and then 381 00:20:32,880 --> 00:20:34,640 Speaker 1: what I want to do in the evening is discharge 382 00:20:34,640 --> 00:20:36,639 Speaker 1: because that's where all the demand does. And if you 383 00:20:36,640 --> 00:20:40,360 Speaker 1: do that with a battery, you'll end up raising demand 384 00:20:40,440 --> 00:20:43,040 Speaker 1: from the rest of the system at midday and lowering 385 00:20:43,080 --> 00:20:46,679 Speaker 1: demand in the evening. And hey, like that helps the 386 00:20:46,720 --> 00:20:49,720 Speaker 1: solar guy. It helps the battery, but that's not the 387 00:20:49,840 --> 00:20:52,240 Speaker 1: end of the story, because it also helps the gas plant. 388 00:20:52,480 --> 00:20:55,320 Speaker 1: Without the battery, there's going to be a gas plant 389 00:20:55,320 --> 00:20:58,560 Speaker 1: somewhere in California, which is watching demand fall at midday, 390 00:20:58,880 --> 00:21:00,560 Speaker 1: they're like, do we turn off? We aren't turn off. 391 00:21:00,600 --> 00:21:05,040 Speaker 1: There's like thermodynamic properties inherent to that gas plant that 392 00:21:05,119 --> 00:21:07,920 Speaker 1: means it's very difficult to switch off and on instantly. 393 00:21:07,960 --> 00:21:10,200 Speaker 1: They can't really do it at all. So they're they're 394 00:21:10,240 --> 00:21:11,960 Speaker 1: in a lot of pain at mid day. So they're 395 00:21:12,000 --> 00:21:15,920 Speaker 1: turning their plant right down to the like minimal technical 396 00:21:16,000 --> 00:21:20,560 Speaker 1: feasible limits they can, hoping that like demand doesn't go 397 00:21:20,600 --> 00:21:22,600 Speaker 1: any lower and they have to turn off, and then 398 00:21:22,640 --> 00:21:25,200 Speaker 1: they breathe this huge sigh of relief when demand comes 399 00:21:25,240 --> 00:21:27,680 Speaker 1: back in the evening and they can turn their plant 400 00:21:27,680 --> 00:21:29,480 Speaker 1: back up and and make their money. And when I 401 00:21:29,520 --> 00:21:32,280 Speaker 1: had batteries to the system, they don't have to turn 402 00:21:32,359 --> 00:21:34,400 Speaker 1: down during the midday, or if I had enough batteries 403 00:21:34,440 --> 00:21:36,399 Speaker 1: and they don't, they get to sit there and operate 404 00:21:36,440 --> 00:21:39,639 Speaker 1: the day through because the batteries are taking the power 405 00:21:39,720 --> 00:21:44,159 Speaker 1: from produced both solar and thereby decreasing overall demand. Is 406 00:21:44,200 --> 00:21:46,800 Speaker 1: that right during the midday, Well, it's it's it's all 407 00:21:46,920 --> 00:21:48,399 Speaker 1: like if you if you're a gas plant, what you 408 00:21:48,400 --> 00:21:51,560 Speaker 1: get paid over the course of that day is very 409 00:21:51,640 --> 00:21:55,159 Speaker 1: very low at midday because demand or net demand so 410 00:21:55,240 --> 00:21:58,240 Speaker 1: low and high in the evening, and if I increased 411 00:21:58,240 --> 00:22:00,399 Speaker 1: demand at midday, then the price goes up and I 412 00:22:00,440 --> 00:22:02,760 Speaker 1: get paid more in midday. I might also get paid 413 00:22:02,920 --> 00:22:05,320 Speaker 1: less in the evening. But like the gas plants will 414 00:22:05,400 --> 00:22:09,080 Speaker 1: have extra additional costs if they're constantly turning their plan 415 00:22:09,240 --> 00:22:11,680 Speaker 1: up and down all the time to sort of make 416 00:22:11,760 --> 00:22:14,240 Speaker 1: room for the rest of the grid which is going crazy. 417 00:22:14,359 --> 00:22:16,560 Speaker 1: So what is optimal? Is it optimal to have a 418 00:22:16,680 --> 00:22:20,120 Speaker 1: bit more expensive batteries or what it's It's a difficult 419 00:22:20,200 --> 00:22:22,919 Speaker 1: question to ask, and that the approach we've taken is 420 00:22:23,400 --> 00:22:26,960 Speaker 1: perhaps a little bit different. So you can write out 421 00:22:27,080 --> 00:22:31,440 Speaker 1: all the equations for this behavior in a specific set 422 00:22:31,440 --> 00:22:33,840 Speaker 1: of instructions and feed it to a computer and get 423 00:22:33,840 --> 00:22:37,320 Speaker 1: a result. It takes an incredible amount of time, but 424 00:22:37,400 --> 00:22:41,399 Speaker 1: you can fully optimize that problem. But we've got slightly 425 00:22:41,440 --> 00:22:44,760 Speaker 1: different priorities when we're producing NEO, because we're not interested 426 00:22:44,840 --> 00:22:51,359 Speaker 1: in a fully optimized, guaranteed perfect dispatch for a system. 427 00:22:51,480 --> 00:22:53,959 Speaker 1: We're interested in in the larger picture. We're interested in 428 00:22:54,000 --> 00:22:57,120 Speaker 1: comparing systems. To make sense to this, I think it's 429 00:22:57,560 --> 00:23:00,640 Speaker 1: maybe maybe useful to borrow one of the the strangers 430 00:23:00,640 --> 00:23:03,960 Speaker 1: alternate realities. For a moment, there's an alternate reality where 431 00:23:04,320 --> 00:23:07,400 Speaker 1: where we've already solved the energy transition, things have gone great. 432 00:23:08,080 --> 00:23:11,240 Speaker 1: Being if still a company that's good. And in this 433 00:23:11,400 --> 00:23:15,440 Speaker 1: sort of zero carbon utopia, people have all these new 434 00:23:15,440 --> 00:23:17,479 Speaker 1: and interesting hobbies, have got all the spare time, and 435 00:23:17,480 --> 00:23:21,000 Speaker 1: bird watching is huge. So imagine that in this alternate reality, 436 00:23:21,240 --> 00:23:24,040 Speaker 1: the B and b n F stands for birds. Birds. 437 00:23:24,320 --> 00:23:27,840 Speaker 1: Birds are big business. Everyone loves birds. And the CEO 438 00:23:27,880 --> 00:23:31,119 Speaker 1: of BENF comes down and he's like, man, people just 439 00:23:31,240 --> 00:23:34,080 Speaker 1: love this bird. This bird content we're producing. We need more. 440 00:23:34,119 --> 00:23:36,840 Speaker 1: And he goes to the best analyst and he's like, 441 00:23:37,280 --> 00:23:40,920 Speaker 1: you have you seen those swarms of starlings. That's sort 442 00:23:40,960 --> 00:23:45,040 Speaker 1: of clouds that undulate and pulse pulsate across the sort 443 00:23:45,040 --> 00:23:46,960 Speaker 1: of evening sky, like I think most of us have 444 00:23:47,040 --> 00:23:49,440 Speaker 1: seen them, and that it says like, yes, sir, it's 445 00:23:49,520 --> 00:23:52,480 Speaker 1: it's actually a murmuration of starlings. That's the collective now. 446 00:23:52,640 --> 00:23:54,800 Speaker 1: And the CEO is like, you're the man for the job, 447 00:23:55,359 --> 00:23:57,240 Speaker 1: or you're the woman for the job. So this analyst 448 00:23:58,240 --> 00:24:02,040 Speaker 1: goes up north. It's very smart, smart person, and they're 449 00:24:02,040 --> 00:24:06,000 Speaker 1: trying to quantify with a set of equations what this 450 00:24:06,000 --> 00:24:09,040 Speaker 1: flock of birds is doing, and they're looking at the 451 00:24:09,080 --> 00:24:12,920 Speaker 1: whole flock, they're looking at the macro and they're trying 452 00:24:12,960 --> 00:24:15,000 Speaker 1: to figure it out. They've got a PhD in maths, 453 00:24:15,040 --> 00:24:17,040 Speaker 1: but they just can't do it. It's just too complex. 454 00:24:17,160 --> 00:24:19,679 Speaker 1: They come up with sort of rough equations, but it 455 00:24:19,720 --> 00:24:22,399 Speaker 1: doesn't hold in all circumstances and things are just falling 456 00:24:22,440 --> 00:24:24,560 Speaker 1: apart and they're pulling their hair out. And the reason 457 00:24:24,680 --> 00:24:27,080 Speaker 1: is because they've they've taken the wrong approach. So if 458 00:24:27,080 --> 00:24:29,680 Speaker 1: you if you want to model this flock of birds, 459 00:24:29,720 --> 00:24:31,640 Speaker 1: a better way to do it is to zoom in 460 00:24:31,800 --> 00:24:36,600 Speaker 1: on the individual bird. So birds aren't particularly intelligent, like 461 00:24:36,640 --> 00:24:39,240 Speaker 1: the expression like bird bird brain is there is there 462 00:24:39,280 --> 00:24:41,520 Speaker 1: for a reason, like they bump into windows all the time. 463 00:24:42,119 --> 00:24:47,120 Speaker 1: And birds are only actually following a few simple set 464 00:24:47,160 --> 00:24:51,359 Speaker 1: of rules when they're flying in flocks. So the first 465 00:24:51,440 --> 00:24:55,400 Speaker 1: rule is basically like don't bump into my neighbors. That's 466 00:24:55,400 --> 00:24:59,000 Speaker 1: pretty simple. The second rule is sort of go roughly 467 00:24:59,040 --> 00:25:02,439 Speaker 1: in the direction that my neighbors are going in, and 468 00:25:02,480 --> 00:25:05,439 Speaker 1: the third is kind of like gravitate towards the center 469 00:25:05,440 --> 00:25:07,879 Speaker 1: of mass of the flock. And it sounds crazy, but 470 00:25:07,920 --> 00:25:12,000 Speaker 1: if you model those individual birds and you program those 471 00:25:12,040 --> 00:25:17,239 Speaker 1: three rules into them, that amazing flocking behavior which we've 472 00:25:17,280 --> 00:25:20,640 Speaker 1: all seen, sort of springs up out of those rules. 473 00:25:21,119 --> 00:25:24,000 Speaker 1: So it's called emergent behavior. And that's a much better 474 00:25:24,080 --> 00:25:29,360 Speaker 1: way of taking this problem and making it something digestible 475 00:25:29,680 --> 00:25:33,280 Speaker 1: that we can get answers from. And so in our reality, 476 00:25:33,359 --> 00:25:35,280 Speaker 1: we we don't have birds, but all we do is 477 00:25:35,320 --> 00:25:38,240 Speaker 1: we have power plants, and we can sort of program 478 00:25:38,440 --> 00:25:42,040 Speaker 1: these simple set of rules into the power plants and 479 00:25:42,080 --> 00:25:45,000 Speaker 1: then sort of set them free, if you will, to 480 00:25:45,000 --> 00:25:48,040 Speaker 1: to follow their own behavior. And that's how we stitch 481 00:25:48,119 --> 00:25:50,639 Speaker 1: the whole system together, which is different like if we 482 00:25:50,760 --> 00:25:52,600 Speaker 1: top down, if we tried to ask the question like 483 00:25:52,640 --> 00:25:55,840 Speaker 1: what's the optimal mix at twelve o'clock on this given 484 00:25:55,920 --> 00:25:59,240 Speaker 1: day for this given demand condition in California, Like that's 485 00:25:59,240 --> 00:26:01,280 Speaker 1: a very difficult question, Like you can't solve it. It 486 00:26:01,320 --> 00:26:03,960 Speaker 1: just takes a very long time. Our approaches say, well, look, 487 00:26:04,080 --> 00:26:06,480 Speaker 1: if you will, like let the power plants decide for themselves. 488 00:26:06,560 --> 00:26:09,960 Speaker 1: We've given them all the cost and operational data they need. 489 00:26:10,080 --> 00:26:12,760 Speaker 1: They can make that decision for themselves. So if I'm 490 00:26:12,800 --> 00:26:16,720 Speaker 1: just random gas plant, you know in Nevada, or whatever 491 00:26:16,800 --> 00:26:19,639 Speaker 1: I can based on certain conditions that you give me, 492 00:26:20,080 --> 00:26:21,919 Speaker 1: I will make a choice on what to do at 493 00:26:21,960 --> 00:26:24,720 Speaker 1: that given hour. Yeah, like if if the price goes 494 00:26:24,760 --> 00:26:27,639 Speaker 1: too low, demand is too low, like I'll switch off 495 00:26:27,440 --> 00:26:29,879 Speaker 1: if you switch the whole thing together. Essentially, what you 496 00:26:29,960 --> 00:26:34,639 Speaker 1: have a bunch of ants scarring alternate realities, trying to 497 00:26:34,760 --> 00:26:42,040 Speaker 1: choose the cheapest flock of power plants? Simple? What can 498 00:26:42,080 --> 00:26:46,200 Speaker 1: NEHEM do that you haven't yet investigated? So what Nehem 499 00:26:46,240 --> 00:26:49,399 Speaker 1: can do that we haven't fully explored yet is a 500 00:26:49,440 --> 00:26:51,880 Speaker 1: lot more work on on zero carbon. So we've done 501 00:26:51,960 --> 00:26:57,440 Speaker 1: some emissions scenarios and they've been very instructive. But what 502 00:26:57,600 --> 00:27:02,320 Speaker 1: we haven't done yet is eached up everything together into 503 00:27:02,400 --> 00:27:05,760 Speaker 1: a cross sectoral optimization. What I mean by that is 504 00:27:05,800 --> 00:27:10,119 Speaker 1: at the moment we've got a view on decarbonization pathways 505 00:27:10,200 --> 00:27:13,640 Speaker 1: for steel for example, and for other sectors. There will 506 00:27:13,680 --> 00:27:18,520 Speaker 1: be electrification as a decarbonization pathway. So transport is a 507 00:27:18,520 --> 00:27:21,560 Speaker 1: good example of that, where we get more electric vehicles 508 00:27:22,000 --> 00:27:24,480 Speaker 1: as long as our electricity is zero carbon, that's fine, 509 00:27:24,520 --> 00:27:27,640 Speaker 1: but it makes the electricity system bigger, so it makes 510 00:27:27,680 --> 00:27:32,000 Speaker 1: that problem more difficult to solve. So we've done quite 511 00:27:32,040 --> 00:27:35,480 Speaker 1: a way into sort of stitching those together at the moment, 512 00:27:35,880 --> 00:27:41,159 Speaker 1: but not everything and not fully optimized. So you have 513 00:27:41,200 --> 00:27:45,280 Speaker 1: all these countries that are making net zero targets and 514 00:27:46,080 --> 00:27:49,120 Speaker 1: you want to model out what that could actually look 515 00:27:49,160 --> 00:27:52,160 Speaker 1: like in practice exactly. And I mean there's a there's 516 00:27:52,160 --> 00:27:55,119 Speaker 1: a difference here as well, because we can take the 517 00:27:55,200 --> 00:27:58,760 Speaker 1: sort of self nominated targets, sort of particular countries done 518 00:27:59,280 --> 00:28:02,439 Speaker 1: and model lat That's a very different question from what's 519 00:28:03,080 --> 00:28:08,479 Speaker 1: the global optimized least cost pathway to zero carbon system? 520 00:28:08,600 --> 00:28:11,680 Speaker 1: Because it might mean some other countries we haven't thought 521 00:28:11,680 --> 00:28:16,160 Speaker 1: of do more heavy lifting than others. It's a it's 522 00:28:16,160 --> 00:28:19,200 Speaker 1: a different story as a whole, rather than the individual 523 00:28:19,240 --> 00:28:21,800 Speaker 1: country story stitched together. One final question I guess is 524 00:28:21,840 --> 00:28:23,359 Speaker 1: what would you like to change on the model for 525 00:28:23,440 --> 00:28:26,360 Speaker 1: next year? Definitely the name. I nominated a bunch of them, 526 00:28:26,400 --> 00:28:28,840 Speaker 1: but they were all shot down. Most of them weren't 527 00:28:28,840 --> 00:28:35,679 Speaker 1: serious conversations Global Renewable Energy Transition Analyzer. So the acronyms GRETTA, 528 00:28:35,760 --> 00:28:39,280 Speaker 1: which I think is very appropriate. Oh man, that's gold. 529 00:28:40,480 --> 00:28:42,480 Speaker 1: I love it. Why would you not do that? That's 530 00:28:42,720 --> 00:28:47,000 Speaker 1: that's great, that's so good. Ian, Thanks for joining us, 531 00:28:47,040 --> 00:28:57,080 Speaker 1: Thanks for having me Today's episode of Switched On was 532 00:28:57,160 --> 00:29:00,200 Speaker 1: edited by Rex Warner of Great Stoke Media. Bloomberg Guny 533 00:29:00,240 --> 00:29:02,560 Speaker 1: App is a service provided by Bloomberg Finance LP and 534 00:29:02,600 --> 00:29:05,520 Speaker 1: its affiliates. This recording does not constitute, nor should it 535 00:29:05,560 --> 00:29:09,480 Speaker 1: be construed as investment advice, investment recommendations, or a recommendation 536 00:29:09,560 --> 00:29:12,600 Speaker 1: as to an investment or other strategy. Bloomberginn EPP should 537 00:29:12,640 --> 00:29:15,240 Speaker 1: not be considered as information sufficient upon which to base 538 00:29:15,280 --> 00:29:18,800 Speaker 1: an investment decision. Neither Bloomberg Finance LP nor any of 539 00:29:18,800 --> 00:29:21,800 Speaker 1: its affiliates makes any representation or warranty as to the 540 00:29:21,840 --> 00:29:25,080 Speaker 1: accuracy or completeness of the information contained in this recording, 541 00:29:25,160 --> 00:29:27,520 Speaker 1: and any liability as a result of this recording. Did 542 00:29:27,520 --> 00:29:28,360 Speaker 1: expressly disclose