WEBVTT 00:02.850 --> 00:04.710 So, I would like to continue. 00:06.150 --> 00:13.310 The last lecture was finished with the last part of that self 00:13.310 --> 00:16.030 -organized balancing, load balancing. 00:16.550 --> 00:22.730 And now we get to a topic which is a very active area of research here 00:22.730 --> 00:23.350 in our group. 00:23.970 --> 00:25.670 Smart home and energy management. 00:26.790 --> 00:29.930 Again, having to do with decentralized systems. 00:30.650 --> 00:35.250 And as you know, we have to find out how we can actually modify the 00:35.250 --> 00:36.630 energy schedules locally. 00:37.350 --> 00:40.670 And so, this is something we looked at intensively. 00:42.270 --> 00:46.950 And there have been quite a few research programs in Germany on that. 00:46.950 --> 01:00.950 One was the so-called E-Energy Program of the Federal Government from 01:00.950 --> 01:02.690 2008 to 2012. 01:03.450 --> 01:08.270 We had six model regions and I already told you about that. 01:08.730 --> 01:13.430 The idea was to combine energy technology with market mechanisms and 01:13.430 --> 01:13.850 ICT. 01:15.610 --> 01:20.650 There, the most important objective was to improve the efficiency of 01:20.650 --> 01:23.590 the energy system and reduce greenhouse gas emissions. 01:24.270 --> 01:27.850 Meanwhile, we know flexibility is one of the major objectives. 01:29.170 --> 01:32.550 And then we had the research programs in the economic incentive 01:32.550 --> 01:36.130 package running from 2009 to 2011. 01:36.130 --> 01:42.450 And we had programs like ICT for electric mobility, like how can we 01:42.450 --> 01:47.130 integrate intelligently the electricity needs of electric vehicles. 01:47.730 --> 01:50.030 We had eight model regions for that. 01:50.370 --> 01:53.170 And we also had large programs all of a sudden on electric storage. 01:53.410 --> 01:58.410 You must be aware of the fact that before that time, like several 01:58.410 --> 02:02.430 years before that, all the funding programs for research on batteries 02:02.430 --> 02:03.070 were stopped. 02:03.070 --> 02:07.730 There was not seen any reason why we should invest in batteries. 02:08.270 --> 02:12.550 All the production facilities, the companies in Germany, we had strong 02:12.550 --> 02:15.930 industry in the battery area. 02:16.450 --> 02:17.330 There was no business. 02:18.430 --> 02:23.510 And so this business moved somewhere else, Korea and somewhere else. 02:23.510 --> 02:27.530 They did work on that. 02:28.050 --> 02:38.630 Certainly for all kinds of devices in MP4 players or all kinds of 02:38.630 --> 02:41.130 entertainment systems, you needed batteries. 02:42.270 --> 02:45.410 But in Germany, it was not seen as an important area. 02:45.770 --> 02:49.130 And so all of a sudden, people noticed, oh, we need batteries. 02:49.630 --> 02:53.810 And then this research was started again. 02:54.950 --> 02:56.770 And Karlsruhe was quite active there. 02:57.550 --> 03:01.210 And I will show you just a little bit about two projects that we made. 03:01.310 --> 03:03.750 A little bit more details than what I showed you before. 03:04.710 --> 03:09.130 Meragio and Meragio Mobile, which were extended with Chrome and Isoys. 03:09.270 --> 03:11.070 But I will just show a little bit about that. 03:11.150 --> 03:14.810 So this was the picture of that funding program. 03:14.810 --> 03:18.210 Six model regions spread over Germany. 03:19.090 --> 03:20.830 Meragio here at the bottom. 03:21.870 --> 03:29.230 And then those associated e-mobility projects, where we again were 03:29.230 --> 03:32.190 active in the Meragio Mobile project. 03:33.230 --> 03:36.210 And I just want to tell you a little bit about Meragio, what we did 03:36.210 --> 03:36.530 there. 03:37.810 --> 03:42.690 So you have seen the slide already in the introductory lecture. 03:43.530 --> 03:46.810 So I don't have to tell you more about that in a moment. 03:47.310 --> 03:50.630 What was the content of that project? 03:51.310 --> 03:57.990 It started with involving just rather few customers. 03:58.750 --> 04:01.190 Just 100 households were involved there. 04:01.570 --> 04:05.690 Where they were confronted with a certain Stromampel. 04:06.630 --> 04:10.850 So a device which could show in different colors, different prices. 04:10.850 --> 04:15.630 And people would respond to those prices and modify their power 04:15.630 --> 04:16.130 consumption. 04:17.130 --> 04:24.710 Then we had actual control schemes where individual devices in 04:24.710 --> 04:27.350 households were controlled by external signals. 04:28.270 --> 04:30.090 So some external optimization. 04:31.190 --> 04:37.670 And finally 40 households were equipped with batteries. 04:37.670 --> 04:42.050 And then those batteries could also be used in order to make better 04:42.050 --> 04:43.930 use of energy at lower prices. 04:44.770 --> 04:48.430 And finally the people were confronted with market mechanisms. 04:49.210 --> 04:53.690 Essentially they could participate in some kind of game, which was 04:53.690 --> 04:57.490 similar to a market, where they could deal with certain energy 04:57.490 --> 05:00.910 products with their own energy consumption and things like that. 05:01.910 --> 05:06.290 So this was a very interesting project where we cooperated with 05:06.290 --> 05:08.910 different companies in the region. 05:09.450 --> 05:17.490 And we had to see how users actually respond to those devices which 05:17.490 --> 05:21.290 show the current power price, the current tariff. 05:21.710 --> 05:23.790 Here in yellow is a medium price. 05:23.930 --> 05:28.370 That's the current power price, a medium value. 05:28.370 --> 05:35.010 And you see the complete schedule for the day, which shows that in the 05:35.010 --> 05:38.630 night hours there is a low price tariff, a green tariff. 05:39.330 --> 05:42.530 And then there are some steps with the yellow color. 05:43.070 --> 05:48.010 And there are these high priced times, sometimes in the day. 05:48.350 --> 05:52.790 And you see that in advance, and so you can in some way plan how you 05:52.790 --> 05:53.810 would use your devices. 05:53.810 --> 05:55.230 That was the idea. 05:55.830 --> 06:00.930 Give some information on the current situation and also for the next 06:00.930 --> 06:01.410 hours. 06:01.630 --> 06:04.630 And in that way provide information for changing the behavior. 06:05.230 --> 06:08.030 How can you measure the changes in behavior? 06:08.330 --> 06:13.990 You need a reference group, which is not doing anything like that. 06:14.690 --> 06:17.790 And you have to compare whether they are more or less the same. 06:17.790 --> 06:23.710 So there was a measurement over some period of time, that's a week, 06:24.850 --> 06:28.650 where people would be looked at in the power consumption. 06:28.870 --> 06:41.270 And so here you have in green and red the group in Mirageo, here in 06:41.270 --> 06:44.250 red, and the reference in green, the reference group. 06:44.250 --> 06:50.130 And then people were confronted with that technology, and you could 06:50.130 --> 06:51.370 see certain changes. 06:51.690 --> 06:55.030 So here you see that the red... no, here now it's different. 06:55.670 --> 07:01.430 Here in green you see the reference, and in red you see the demand for 07:01.430 --> 07:04.750 the group from the Mirageo test group. 07:05.150 --> 07:09.510 And you see that now there are quite a few differences in power 07:09.510 --> 07:10.470 consumption. 07:10.470 --> 07:16.070 And this can be now looked at in more details, looking at the 07:16.070 --> 07:17.430 different times of the day. 07:17.750 --> 07:24.130 And different times of the day, all the different prices have been 07:24.130 --> 07:27.130 shown to the users at different days. 07:29.250 --> 07:33.350 So then you could see how they would respond to that, and this is 07:33.350 --> 07:35.490 shown in more detail in this slide. 07:35.490 --> 07:41.870 So over a longer period of time, again, compared to a reference group, 07:42.270 --> 07:46.790 how would the people actually change their power consumption? 07:47.470 --> 07:53.130 And here we only show the green and the red curve. 07:53.290 --> 07:58.110 So green for low price, red for high price. 07:58.110 --> 08:07.230 And for... so if we see here 24%, that's the peak value in winter 08:07.230 --> 08:09.850 actually, the peak value in summer was even higher. 08:10.570 --> 08:17.030 It means that the power consumption was increased by 25% in summer 08:17.030 --> 08:20.510 here at 3pm in the afternoon. 08:22.610 --> 08:28.970 And actually in the morning, you see that in the winter, it was 08:28.970 --> 08:30.910 always... 08:30.910 --> 08:36.210 the power consumption was decreased at those times. 08:38.210 --> 08:42.410 And so for the red curve, you certainly always assume that it should 08:42.410 --> 08:47.770 be in the negative part, because we would like to decrease the power 08:47.770 --> 08:48.350 consumption. 08:48.350 --> 08:53.010 But you see that it's not always the case that it works. 08:53.130 --> 08:57.610 There are certain special effects at that time, and here also in the 08:57.610 --> 08:57.930 morning. 08:58.810 --> 09:02.350 And during the day, you have quite some flexibility, which is there 09:02.350 --> 09:07.710 just by manual response to those price signals. 09:08.310 --> 09:14.610 And it shows that there is quite a potential for load shifting during 09:14.610 --> 09:15.030 the day. 09:15.730 --> 09:22.830 And it is obvious that if you would have an automated response and 09:22.830 --> 09:26.170 more complete information on the potential flexibility of your 09:26.170 --> 09:34.610 devices, you could even get more flexibility of demand. 09:36.050 --> 09:40.930 So that is the load flexibility that we observed in the pilot region 09:40.930 --> 09:41.880 with 1000 customers. 09:42.950 --> 09:48.970 And the major insight is that there is quite a potential for load 09:48.970 --> 09:51.850 shifting during the day in households. 09:52.750 --> 09:54.950 If you have many households, you have an effect. 09:55.110 --> 09:57.850 If you have one household, it's a small effect. 09:57.850 --> 10:02.150 Very often people argue and say, oh, a fridge cannot do anything. 10:02.730 --> 10:07.170 If you have 2000 fridges, they can do something. 10:07.650 --> 10:11.650 If they would work in synchrony, it would be very bad. 10:12.170 --> 10:13.290 You don't want that. 10:13.710 --> 10:17.550 But nevertheless, you can do something with those devices. 10:20.150 --> 10:22.390 We did a few more things. 10:22.870 --> 10:23.470 Definitely Mirageo. 10:23.590 --> 10:24.570 I will come back to that. 10:24.730 --> 10:29.970 In particular, we designed a certain feedback cycle between the 10:29.970 --> 10:33.210 households and the grid operator. 10:34.270 --> 10:39.530 And designed a way of actually responding or getting a good demand 10:39.530 --> 10:43.850 response in those areas. 10:43.850 --> 10:48.990 Then we had Mirageo Mobile, where the major point was to integrate the 10:48.990 --> 10:52.450 electric vehicles or the energy demand into the smart grid. 10:54.330 --> 10:58.410 I told you already about this slide here. 10:59.510 --> 11:05.130 And you know that we have a certain German development plan for 11:05.130 --> 11:05.930 electric mobility. 11:05.930 --> 11:11.830 So our project was running in that first phase. 11:12.310 --> 11:15.290 Now we are in the second phase, the market development. 11:16.110 --> 11:23.090 The idea is that in 2020, in five years, we should have around one 11:23.090 --> 11:24.430 million electric vehicles. 11:24.430 --> 11:27.650 Meanwhile, people say not just electric vehicles. 11:28.950 --> 11:38.390 It can be a battery electric vehicle or it could be a plugged hybrid 11:38.390 --> 11:39.870 electric vehicle. 11:39.870 --> 11:42.150 And then this looks more realistic. 11:43.010 --> 11:50.530 Because the plugged hybrid vehicles, they can actually also drive with 11:50.530 --> 11:57.050 extra combustion engine and then it sounds more realistic. 11:57.270 --> 12:00.810 But people are not convinced that we will get that. 12:01.190 --> 12:05.110 The idea is that in 2030, we will get six million electric vehicles. 12:05.110 --> 12:07.850 Maybe the case, you don't know. 12:09.270 --> 12:14.830 You may have observed that in Germany, we actually, like our major car 12:14.830 --> 12:17.610 companies, they do have electric vehicles. 12:18.230 --> 12:25.990 But they are not really making that much or pushing that much to 12:25.990 --> 12:27.590 actually use electric vehicles. 12:27.590 --> 12:31.250 There are other companies which are quite more active in that area. 12:32.030 --> 12:36.070 Although we are also quite active, we have the leading edge cluster of 12:36.070 --> 12:37.590 electric mobility in the southwest. 12:38.730 --> 12:48.990 But our large automotive companies are a bit reluctant to have real 12:48.990 --> 12:51.890 large rollout of electric vehicles. 12:51.890 --> 12:56.950 Certainly, the battery technology is not at a stage that we would like 12:56.950 --> 12:57.450 it to be. 12:57.950 --> 13:01.890 So batteries are still very expensive and this is a major obstacle to 13:01.890 --> 13:06.130 more production of electric vehicles. 13:06.490 --> 13:12.750 We also don't have an industrialized production facility for batteries 13:12.750 --> 13:13.430 in Germany. 13:14.390 --> 13:16.890 Tesla is building one in North America. 13:17.330 --> 13:19.710 In Asia, we have production facilities. 13:19.870 --> 13:24.070 In Germany, we don't have a production facility, which is a bad thing. 13:25.030 --> 13:28.210 But people are working on battery technology. 13:28.210 --> 13:36.630 Okay, so the idea is to have that development for several reasons. 13:37.070 --> 13:41.830 For environmental reasons, definitely, because if you have an electric 13:41.830 --> 13:47.430 vehicle, you don't produce carbon dioxide while using the vehicle. 13:47.430 --> 13:48.830 Certainly, you use power. 13:49.910 --> 13:52.670 And when you use power, you have to look at how that power is 13:52.670 --> 13:53.790 generated, electricity. 13:54.570 --> 13:58.930 If you have conventionally generated electricity, then you have a 13:58.930 --> 14:03.470 certain mixture of different fossil fuels that are used. 14:03.570 --> 14:07.750 And so you have carbon dioxide emission for the generation of the 14:07.750 --> 14:09.310 power that you are using for driving. 14:09.310 --> 14:15.690 If you are just using power from your rooftop, it may be actually zero 14:15.690 --> 14:16.090 emission. 14:16.810 --> 14:19.590 But this depends on what you're actually doing, how you're operating 14:19.590 --> 14:20.210 your vehicle. 14:22.230 --> 14:25.650 Now, why are actually these electric vehicles of interest? 14:25.790 --> 14:31.390 One point is we would like to have less carbon dioxide emissions, but 14:31.390 --> 14:34.010 there are more interesting effects that we can see there. 14:34.010 --> 14:38.430 So, let's look at the typical characteristics of electric vehicles. 14:38.970 --> 14:45.470 So, if you look at the mobility in Germany, there's a survey, a 14:45.470 --> 14:49.310 regular survey, taken on the mobility of German citizens. 14:50.470 --> 14:53.810 Taken, actually, by an institute here at KIT. 14:54.970 --> 15:01.210 And there's the so-called mobility panel, and there it is obvious 15:01.210 --> 15:06.410 that, well, the average daily car usage is less than an hour. 15:07.130 --> 15:11.310 You also know that from your own experience, usually you don't drive 15:11.310 --> 15:12.590 more than an hour per day. 15:13.150 --> 15:16.050 Maybe half an hour to work and half an hour back. 15:17.550 --> 15:21.630 And 94% of our trips are less than 50 kilometers. 15:21.630 --> 15:28.230 That means that if you have a vehicle which can drive 100 kilometers 15:28.230 --> 15:32.670 easily, it's perfect for urban traffic. 15:33.070 --> 15:37.750 Because you don't drive that much, the car is sitting around most of 15:37.750 --> 15:38.490 the time. 15:38.490 --> 15:43.010 Sitting somewhere on a parking lot in your garage or wherever. 15:44.570 --> 15:48.830 And it means there's only a short time where the car is not parked 15:48.830 --> 15:49.370 somewhere. 15:50.090 --> 15:55.290 If it's parked somewhere, it might be within reach of the electricity 15:55.290 --> 15:55.910 system. 15:56.570 --> 15:58.370 Or the power grid. 15:59.050 --> 16:02.970 And so most of the time, these vehicles could be connected to the 16:02.970 --> 16:03.230 grid. 16:04.950 --> 16:06.150 And now it gets interesting. 16:06.970 --> 16:11.990 Because it's driving for one hour, less than one hour, it can be 16:11.990 --> 16:15.870 connected to the grid most of the time, and then you have, most of the 16:15.870 --> 16:19.070 time, you have a battery that is connected to the grid. 16:19.790 --> 16:23.130 Now it gets interesting to see what we can do with that. 16:24.210 --> 16:29.690 So, the currently available electric vehicles have batteries with a 16:29.690 --> 16:32.330 capacity around 20 kilowatt hours. 16:33.290 --> 16:35.250 Some have more, some have less. 16:35.870 --> 16:47.810 If you have a Tesla S, if it's well equipped, you have 85 kilowatt 16:47.810 --> 16:48.190 hours. 16:49.170 --> 16:55.330 So that's definitely quite large, much larger than the other ones. 16:55.430 --> 16:59.670 But the normal vehicles that we have, have around 20 kilowatt hours, 16:59.770 --> 17:07.110 which means that you can, with that you can go for around 100 to 150 17:07.110 --> 17:07.690 kilometers. 17:10.270 --> 17:12.490 That's the typical value. 17:13.270 --> 17:14.230 Now what does this mean? 17:14.290 --> 17:20.130 If we have one million battery electric vehicles, that's a storage 17:20.130 --> 17:22.150 capacity of 20 gigawatt hours. 17:22.150 --> 17:28.930 And if now these batteries are connected to the network, for example, 17:29.170 --> 17:38.490 single phase connected means at 230 volts and 16 ampere means 3.7 17:38.490 --> 17:40.630 gigawatt potential power. 17:41.950 --> 17:49.750 Like 3.7 kilowatt power for an individual vehicle that is connected to 17:49.750 --> 17:50.150 the grid. 17:50.430 --> 17:55.030 And if you have one million, you have 3.7 gigawatt potential power 17:55.030 --> 18:03.770 that can be provided by those vehicles or if they are recharged in 18:03.770 --> 18:05.530 need of 3.7 gigawatt. 18:06.250 --> 18:07.870 They are all connected to the grid. 18:09.090 --> 18:12.170 Now that's about three nuclear power plants. 18:14.370 --> 18:16.750 So that's just for one million electric vehicles. 18:17.550 --> 18:25.110 So it can be quite some demand, can also be quite some supply, if the 18:25.110 --> 18:28.990 vehicles are capable to actually feed in power into the grid. 18:29.990 --> 18:37.250 In Germany, most cars, car manufacturers don't agree to let the 18:37.250 --> 18:39.710 vehicles provide power into the grid. 18:40.370 --> 18:47.350 They restrict that and they say the battery is only for driving for 18:47.350 --> 18:47.990 nothing else. 18:47.990 --> 18:59.590 In Japan, it is mandatory that every electric vehicle must be able to 18:59.590 --> 19:01.050 feed power into the grid. 19:01.770 --> 19:07.050 Because they have the situation that they have almost no nuclear power 19:07.050 --> 19:07.790 plants anymore. 19:07.990 --> 19:11.490 They had something like around 50 nuclear power plants. 19:11.490 --> 19:13.330 They had all been switched off. 19:13.990 --> 19:16.150 The current government tries to switch them on again. 19:16.690 --> 19:22.430 But nevertheless, they have quite a few bottlenecks in the energy 19:22.430 --> 19:22.930 system. 19:23.570 --> 19:26.970 And so the situation is there quite common that you have electric 19:26.970 --> 19:29.890 vehicles which would power a traffic light. 19:29.890 --> 19:33.310 Things like that are mandatory there. 19:33.950 --> 19:36.990 And so it's a different development of technologies. 19:38.710 --> 19:38.910 Okay. 19:39.730 --> 19:43.290 Consequently, definitely, you have a high demand for power, but 19:43.290 --> 19:46.890 potentially also a high supply if power feedback is possible. 19:47.810 --> 19:48.710 You have a question? 19:49.390 --> 19:49.570 From 19:54.200 --> 19:58.240 where did you get that information? 20:01.000 --> 20:01.240 Yes. 20:03.200 --> 20:04.160 Did you? 20:10.700 --> 20:11.460 Cycles. 20:11.460 --> 20:11.980 Yes. 20:15.980 --> 20:16.860 But 20:28.210 --> 20:38.830 there are different experiences with different technologies for 20:38.830 --> 20:39.170 batteries. 20:40.150 --> 20:44.770 And for example, we had electric vehicles which was provided to us by 20:44.770 --> 20:45.490 Opel. 20:46.430 --> 20:52.690 And they had made many experiments, like test runs with their 20:52.690 --> 20:55.590 batteries because they have the Ampera technology and so on. 20:56.410 --> 21:00.710 And they had not seen aging in their batteries. 21:00.830 --> 21:01.790 Almost no aging. 21:03.230 --> 21:08.150 Certainly, they did not run complete cycles from zero state of charge 21:08.150 --> 21:10.910 to 100% state of charge, but only a smaller fraction. 21:12.070 --> 21:17.350 And if we talk about power feedback, we always only talk about a small 21:17.350 --> 21:22.050 fraction of the state of charge that would be used for power feedback. 21:22.750 --> 21:29.970 And there actually have been some observations that batteries actually 21:29.970 --> 21:44.130 can benefit from being used in a rather mild way by feeding back power 21:44.130 --> 21:44.930 into a house. 21:44.930 --> 21:53.450 Because if you look at the power situation or the way a battery is 21:53.450 --> 21:57.650 used while it's driving, it looks like this. 21:58.350 --> 22:03.070 And if it's recharged, it looks like this. 22:05.030 --> 22:10.950 So, while it's driving, there are many cycles, many short cycles. 22:12.950 --> 22:18.550 It's a very dynamic situation, different voltages, different power 22:18.550 --> 22:19.590 that is needed. 22:20.590 --> 22:23.450 And so, it's quite some stress on the battery while it's driving. 22:24.330 --> 22:29.910 If it's connected to a house, maybe you have one additional cycle or 22:29.910 --> 22:30.690 something like that. 22:30.870 --> 22:35.390 But you don't have a characteristic profile like when it's driving. 22:35.390 --> 22:41.110 And so, this situation that you would use extra if you would use it 22:41.110 --> 22:45.090 for feeding back power once or twice while it's connected to the house 22:45.090 --> 22:48.870 is not really adding to the aging. 22:49.330 --> 22:52.070 The aging is coming in that area. 22:53.350 --> 22:58.930 So, it is a very difficult issue because certainly the car 22:58.930 --> 23:05.510 manufacturers say we have to prevent every additional load on the 23:05.510 --> 23:05.790 battery. 23:06.330 --> 23:10.890 No additional cycle should be used only for driving because we have a 23:10.890 --> 23:12.910 restricted battery lifetime. 23:14.330 --> 23:20.050 But meanwhile, this has changed as far as I know. 23:20.130 --> 23:23.930 I'm not a battery specialist, but as far as I know, things are 23:23.930 --> 23:28.550 changing and they are meanwhile expecting lifetimes, like they give 23:28.550 --> 23:31.230 guarantees for 8 years. 23:31.910 --> 23:37.110 Certainly, it means that if the battery is dead before that, they have 23:37.110 --> 23:38.250 to replace it. 23:38.510 --> 23:44.470 But they dare to give guarantees for battery lifetime of 8 years. 23:45.450 --> 23:48.230 And so, it should be possible. 23:49.850 --> 23:51.630 Technology is improving. 23:52.470 --> 23:56.890 And certainly, if you look at batteries in your devices, laptop, 23:57.150 --> 24:01.530 smartphone and so on, we know that after a year or two years, they 24:01.530 --> 24:03.770 wear out and you have to replace them. 24:06.110 --> 24:10.530 But, like the battery I have here, it lasts longer. 24:11.570 --> 24:14.130 So, people are working on that. 24:15.230 --> 24:20.090 And in particular, like the battery technology, there is quite some 24:20.090 --> 24:22.050 progress has been made in recent years. 24:22.730 --> 24:27.830 And also Tesla would not use that if the battery would not run or 24:27.830 --> 24:31.090 would not last longer than 1 or 2 years. 24:31.270 --> 24:32.790 That would not make sense. 24:33.830 --> 24:35.610 I see the concern. 24:37.010 --> 24:39.270 And it's an important issue, definitely. 24:39.490 --> 24:42.410 So, battery technology is an important issue with respect to that. 24:42.590 --> 24:47.470 If you cannot provide reliable batteries, then it's too expensive. 24:48.870 --> 24:55.490 But it has to look at, even if they are not used for providing power 24:55.490 --> 25:00.730 feedback, do they still have a positive impact on the energy system? 25:01.510 --> 25:03.450 That's what I would like to address here. 25:04.290 --> 25:09.090 So, what we see is that if we look at the time for recharging a 25:09.090 --> 25:15.390 battery, if we have single-phase recharging at 3.7 kilowatt, assuming 25:15.390 --> 25:20.890 that the battery is completely empty, you need 5 to 7 hours. 25:20.890 --> 25:26.470 If you have three-phase charging, it's about 2 hours. 25:27.190 --> 25:31.490 But to get 80%, it can be even faster. 25:32.030 --> 25:35.470 So, for example, that's not charging at home. 25:35.470 --> 25:41.930 The Tesla, if you would use that on your home power outlet, 3.7 25:41.930 --> 25:47.150 kilowatt, you can see how long it would take to get back 85 kilowatt 25:47.150 --> 25:47.490 hours. 25:48.110 --> 25:49.490 That does not make sense. 25:50.950 --> 25:51.970 It's just too long. 25:52.310 --> 25:54.470 It's more than 20 hours. 25:55.290 --> 25:59.710 So, they usually have fast charging at very high power. 26:00.330 --> 26:05.890 They actually can do it at such a high power that they can get up to a 26:05.890 --> 26:08.410 state of charge of 80% within 20 minutes. 26:09.430 --> 26:15.450 Tesla is just building those stations along the major routes in 26:15.450 --> 26:18.090 Germany and certainly in North America. 26:19.190 --> 26:23.530 Okay, this is average time for charging. 26:24.290 --> 26:28.110 But it means that the time that they need for charging is rather 26:28.110 --> 26:30.570 small, just a fraction of the day. 26:31.290 --> 26:34.050 And most of the time, the car is just sitting there. 26:35.020 --> 26:41.290 And so, the time where it's actually being recharged is quite 26:41.290 --> 26:41.750 flexible. 26:43.270 --> 26:52.190 And when you get home, put in your plug or the cable of your vehicle 26:52.190 --> 26:59.210 and at a certain time you put in your plug, you have a certain time 26:59.210 --> 27:01.890 where you would like to start again. 27:02.430 --> 27:06.410 Usually, you should tell the system at what time you would like to 27:06.410 --> 27:10.050 start again, at what time you would like to make the next drive with 27:10.050 --> 27:14.710 the vehicle, because at that point in time, you should have a fully 27:14.710 --> 27:15.530 charged vehicle. 27:16.470 --> 27:21.450 If it would be fully charged for a long time, that would not be good. 27:22.370 --> 27:27.350 If you would start to recharge it immediately and then let it sit 27:27.350 --> 27:33.130 there and it's fully charged, this is contributing to the aging of the 27:33.130 --> 27:33.430 battery. 27:33.430 --> 27:38.310 Like a fully charged battery should be recharged immediately, or at 27:38.310 --> 27:41.730 least it shouldn't take too long before it is discharged. 27:43.030 --> 27:49.990 So, it means that this actual recharging process should be shifted to 27:49.990 --> 27:50.870 some later time. 27:53.010 --> 27:55.450 And now you have all this flexibility. 27:56.550 --> 28:03.090 You can use that for flexibility of consumption, of energy. 28:04.050 --> 28:08.270 And this is the major effect that we have, that we have a high 28:08.270 --> 28:13.350 flexibility for load shifting from electric vehicles. 28:14.350 --> 28:21.250 This certainly depends on the capability to actually influence the 28:21.250 --> 28:25.750 time at which a battery is recharged, but the protocols that are 28:25.750 --> 28:28.490 developed or that have been developed allow for that. 28:29.290 --> 28:35.210 And certainly our current electric vehicles that are on the road, they 28:35.210 --> 28:40.930 would just start recharging at the moment you plug in the cable. 28:41.770 --> 28:47.130 But if you have an intelligent system, then they would have that 28:47.130 --> 28:54.490 protocol and then you could actually control the time for recharging. 28:56.810 --> 29:02.570 So, this is possible in those projects that we have run. 29:02.690 --> 29:06.230 We have developed all those technologies, the protocols for that, and 29:06.230 --> 29:10.590 have done that with the vehicles that were on the road in those 29:10.590 --> 29:11.510 projects. 29:12.410 --> 29:17.110 And now what we have to see is how we can actually use that potential. 29:17.410 --> 29:23.110 So, the idea is that it would be bad if, like this is showing the 29:23.110 --> 29:31.230 consumption curve in black of part of the network of the power grid 29:31.230 --> 29:32.170 with 100 households. 29:32.170 --> 29:36.950 Now, if electric vehicles come home, are all plugged in, when the 29:36.950 --> 29:41.050 people come home, you have a load curve like that indicated here. 29:41.610 --> 29:45.870 And this would be bad to have an extra load at a time where you have a 29:45.870 --> 29:46.950 load peak anyway. 29:46.950 --> 29:53.730 And so, the first idea would be to shift that from the evening hours 29:53.730 --> 29:54.910 to the night hours. 29:55.710 --> 30:00.250 Simple thing because people get home in the evening, they would like 30:00.250 --> 30:03.810 to start in the morning again and it's sufficient that the batteries 30:03.810 --> 30:06.210 are recharged just before they are leaving. 30:07.150 --> 30:12.610 And then you could say, I would like to drive my vehicle with solar 30:12.610 --> 30:13.390 power only. 30:14.290 --> 30:18.030 And then you could use, like here, this is indicating the in-feed of 30:18.030 --> 30:19.550 power from solar panels. 30:20.150 --> 30:23.750 Then you could recharge the battery exactly at that time when you get 30:23.750 --> 30:24.670 power from the rooftop. 30:24.670 --> 30:31.210 And maybe you have the capability to feed back energy to the grid. 30:31.790 --> 30:39.050 You could even out the load peaks of that area and this would result 30:39.050 --> 30:45.330 in, indicated in green here, this is a load curve which is almost 30:45.330 --> 30:45.890 stable. 30:47.550 --> 30:52.450 The goal is not necessarily to have such a load curve which is just a 30:52.450 --> 30:56.410 stable voltage or stable power during the day. 30:56.730 --> 31:01.730 But the essential point is that we can generate a range of different 31:01.730 --> 31:08.650 demand profiles just by using the recharging needs of electric 31:08.650 --> 31:09.210 vehicles. 31:09.210 --> 31:12.630 As long as they are connected all the time with the grid. 31:14.210 --> 31:19.190 So this is what makes the electric vehicles interesting. 31:20.250 --> 31:24.530 Some people say, oh, it's not necessary really to build a car in order 31:24.530 --> 31:26.310 to get a battery into the house. 31:27.270 --> 31:29.190 Just to get a battery is cheaper. 31:30.130 --> 31:34.750 People will do that anyway, but the vehicles will be there also. 31:35.430 --> 31:38.870 And so if they are there, they should be used in the best possible 31:38.870 --> 31:39.150 way. 31:40.070 --> 31:43.730 And so we should satisfy the recharging needs in order to provide the 31:43.730 --> 31:51.290 mobility, but we should do it such that the grid benefits from that 31:51.290 --> 31:52.670 potential flexibility. 31:54.110 --> 31:58.330 And in order to work on that, we have designed this energy smart home 31:58.330 --> 32:08.430 lab at KIT, which here is... this photo was at a time when it was 32:08.430 --> 32:10.570 actually sitting on a parking lot. 32:11.090 --> 32:14.490 On that parking lot, we now have a new building, so it was moved to 32:14.490 --> 32:14.870 the right. 32:15.530 --> 32:19.250 And the PV panels are not on the top at the moment, but will be placed 32:19.250 --> 32:19.850 on it again. 32:21.370 --> 32:27.770 And inside, we have a two-bedroom apartment with all kinds of 32:27.770 --> 32:29.510 appliances in there. 32:29.650 --> 32:33.790 We also have... we had this electric vehicle from Opel. 32:34.730 --> 32:40.490 And I will show you scenarios around that energy smart home lab and 32:40.490 --> 32:42.870 what kind of technologies we actually have designed there. 32:44.990 --> 32:50.010 So, let me just... before I come back to the smart home lab at KIT, 32:50.790 --> 32:52.910 let me tell you something about smart homes. 32:53.290 --> 32:56.350 The notion of a smart home has been around for quite a long time. 32:56.790 --> 32:58.030 We did not invent that. 32:58.510 --> 33:01.150 And usually it's not related to energy. 33:01.730 --> 33:04.450 Usually it's related to home automation. 33:04.450 --> 33:09.930 In a smart home, you have smart, intelligent devices for discovering 33:09.930 --> 33:15.010 when you are approaching the house, discovering when somebody tries to 33:15.010 --> 33:18.870 get into the house who is not authorized to do so. 33:19.310 --> 33:24.050 You would like to have remote switch control for the lighting, for 33:24.050 --> 33:26.770 windows, for coffee machines, and so on. 33:27.390 --> 33:28.590 So, this is home automation. 33:30.130 --> 33:34.150 You would like to have security and safety, smoke alarm, smoke 33:34.150 --> 33:38.470 detection alarm, door communication, then ambient assisted living, 33:39.630 --> 33:43.850 assistance for elderly people, for disabled people, observation of 33:43.850 --> 33:49.690 what people are doing, whether an old, sick person is falling out of 33:49.690 --> 33:50.610 the bed, and so on. 33:50.610 --> 33:54.890 If you are interested in scenarios like that, come to the House of 33:54.890 --> 33:56.310 Living Labs of FZI. 33:57.030 --> 34:00.510 There we have a scenario like that that you can actually look at and 34:00.510 --> 34:01.870 see how that works. 34:01.870 --> 34:08.850 And then we have all kinds of multimedia devices, so certain displays, 34:08.990 --> 34:14.470 glanceable displays, where you just look at a device and you see, you 34:14.470 --> 34:16.390 get information on what's happening there. 34:16.950 --> 34:22.270 You have music control, email access, all kinds of devices which are 34:22.270 --> 34:24.290 controlled in a nice way. 34:24.690 --> 34:27.270 You get always the environment that you like. 34:27.590 --> 34:30.290 The house detects what you are doing and things like that. 34:30.290 --> 34:34.810 Nice, smart home environment, has nothing to do with energy. 34:34.950 --> 34:38.750 Actually, these things here are energy expensive. 34:39.910 --> 34:45.230 If you have a home automation system in your house, typical technology 34:45.230 --> 34:47.610 needs quite a bit of energy. 34:49.350 --> 34:54.170 Now, what we are putting there now as an additional thing is the 34:54.170 --> 34:55.070 energy aspect. 34:56.010 --> 35:00.830 And meanwhile, if you look into literature on smart homes, most people 35:00.830 --> 35:01.710 talk about energy. 35:02.210 --> 35:07.190 Don't talk that much anymore about comfort, security, and health, but 35:07.190 --> 35:08.270 that's also very important. 35:10.210 --> 35:14.630 So, what can you do if you look at energy aspects in such a house? 35:15.390 --> 35:18.190 You can look at smart metering technology. 35:18.410 --> 35:23.390 We will come back to smart metering in one of the later chapters, how 35:23.390 --> 35:26.250 we can deal with that. 35:26.250 --> 35:32.010 So, smart metering means we have power reading at quite a high 35:32.010 --> 35:33.490 resolution. 35:34.650 --> 35:37.550 We can control the devices. 35:38.330 --> 35:39.870 We have some external signals. 35:39.990 --> 35:43.330 That was what I showed you with respect to the Meragio project. 35:44.650 --> 35:49.190 Then, we have all kinds of interdependencies between households and 35:49.190 --> 35:54.790 distribution grids in order to even out what's happening in the 35:54.790 --> 35:55.770 distribution grid. 35:56.890 --> 36:01.890 We have decentralized power generation, not just photovoltaic panels 36:01.890 --> 36:04.350 on the rooftop, but also combined heat and power plants. 36:05.310 --> 36:10.370 We have all kinds of domestic appliances, which might be smart to some 36:10.370 --> 36:10.710 degree. 36:11.390 --> 36:16.970 So, washing machines, dryers, dishwashers, and so on, but also air 36:16.970 --> 36:19.650 conditioning, deep freezers, fridges. 36:20.430 --> 36:25.450 And we have the electric mobility devices, and then we have all kinds 36:25.450 --> 36:29.310 of optimization goals, objectives to minimize energy consumption. 36:29.410 --> 36:30.750 That's one important objective. 36:30.970 --> 36:32.670 We don't want to pay that much for energy. 36:34.350 --> 36:37.530 So, it would be good to minimize energy consumption. 36:38.210 --> 36:39.910 Minimize carbon dioxide emissions. 36:40.070 --> 36:43.870 If we only have power from the rooftop, we don't have to save on that. 36:43.990 --> 36:44.590 It's just there. 36:45.270 --> 36:46.250 We can consume that. 36:47.550 --> 36:50.070 Maximize self-consumption of photovoltaic generation. 36:50.070 --> 36:54.850 This is an important issue, and this is the major point why people 36:54.850 --> 36:59.730 will buy batteries for stationary use in their houses, because then 36:59.730 --> 37:05.390 you can utilize your PV power from the rooftop at a price that is much 37:05.390 --> 37:08.890 smaller than the price from the utility company. 37:08.890 --> 37:11.810 So, this is something which will be done by the market. 37:12.650 --> 37:15.970 You don't need any subsidies for that. 37:16.310 --> 37:19.810 This just will develop by itself. 37:21.090 --> 37:24.230 And then here I have indicated all kinds of devices. 37:24.450 --> 37:30.510 These are the energy prices that are indicated on those Stromampel 37:30.510 --> 37:31.450 devices here. 37:31.450 --> 37:39.890 This is some app showing energy prices, giving some information on 37:39.890 --> 37:42.530 that Stromampel in a different scenario. 37:44.350 --> 37:47.550 Now, what do we have in our energy smart home lab? 37:48.230 --> 37:50.570 We have intelligent appliances. 37:51.310 --> 37:58.710 Dishwasher, washing machine, dryer, also stove and coffee machine. 37:59.830 --> 38:03.050 Stove and a coffee machine are not that flexible, but a dishwasher 38:03.050 --> 38:03.950 certainly is. 38:04.650 --> 38:06.530 We have smart metering systems. 38:06.870 --> 38:10.790 We certainly have a solar inverter for the PV panels on the rooftop. 38:11.570 --> 38:15.070 We have air conditioning. 38:16.370 --> 38:21.370 Actually, we have a phase change material in the ceiling of the rooms. 38:21.970 --> 38:26.470 And this phase change material is melting when it's getting warm, and 38:26.470 --> 38:29.450 it is crystallizing when it's cold. 38:29.570 --> 38:33.450 So, it has to be made cold actively, so it's cooled down. 38:34.110 --> 38:37.490 And then at some later time, when it's actually getting warm, the 38:37.490 --> 38:38.510 cooling effect is there. 38:39.380 --> 38:41.590 We have a combined heat and power plant. 38:41.990 --> 38:48.190 We also have an electric heater, which means that here we have some 38:48.190 --> 38:51.110 bivalent technology. 38:51.930 --> 38:56.430 You remember that in the beginning I told you about the scenario that 38:56.430 --> 39:01.350 we should be able to transform power into gas, for example, in a 39:01.350 --> 39:06.290 virtual way by switching between use of power and use of gas to 39:06.290 --> 39:07.130 provide heat. 39:07.130 --> 39:09.410 This is exactly what we can do there. 39:09.930 --> 39:13.050 Switch between electric heater and combined heat and power plant to 39:13.050 --> 39:14.030 produce warm water. 39:15.810 --> 39:19.130 Then we have a charging station outside, which can recharge the 39:19.130 --> 39:23.530 battery and can also feed in power into the house. 39:23.670 --> 39:27.790 So, this was possible with the vehicle we had here from Opel. 39:27.790 --> 39:32.390 The only electric vehicle which in an intelligent way can actually 39:32.390 --> 39:35.490 reach or can feed back power into the house. 39:35.890 --> 39:40.110 No other company was capable to do that, only Opel, in cooperation 39:40.110 --> 39:40.590 with us. 39:41.410 --> 39:43.550 The other companies did not do that. 39:43.630 --> 39:47.130 They promised to do it, but they did not provide a vehicle that could 39:47.130 --> 39:47.650 do that. 39:48.310 --> 39:54.030 And then we have devices to inform the residents of the house about 39:54.030 --> 39:55.470 the current energy situation. 39:55.750 --> 39:59.430 Something which we call an energy management panel, which is providing 39:59.430 --> 40:01.290 information of what's happening in the house. 40:02.190 --> 40:05.270 I will give you information on that shortly to you directly. 40:07.230 --> 40:11.510 And we have an energy management system, which uses all the 40:11.510 --> 40:15.530 information about the devices, about the energy generation, energy 40:15.530 --> 40:20.490 consumption and the preferences of the user, which are specified using 40:20.490 --> 40:23.890 those energy management panel menus. 40:24.650 --> 40:30.030 That information is used in the house to optimize the schedule for the 40:30.030 --> 40:30.250 house. 40:30.310 --> 40:36.510 So, this is one central optimization device inside the house, which is 40:36.510 --> 40:42.450 optimizing the schedule for the house. 40:43.870 --> 40:46.690 And then we have even more devices. 40:47.130 --> 40:50.690 We have a so-called PV power simulator. 40:51.530 --> 40:54.190 Very important, because this is a lab. 40:54.590 --> 40:56.150 We would like to run experiments. 40:56.150 --> 41:04.290 We would like to be able to actually put in a situation or physically 41:04.290 --> 41:09.330 simulate a situation as if we would have power generation from the 41:09.330 --> 41:13.730 rooftop, but this is just a replay of what has been measured before, a 41:13.730 --> 41:14.790 certain power profile. 41:14.790 --> 41:19.850 And then this can be replayed exactly like that, or it could be scaled 41:19.850 --> 41:20.790 in some way. 41:21.710 --> 41:25.770 And in this way, we can make experiments in how the house actually 41:25.770 --> 41:31.690 would respond to a certain in-feed of solar power. 41:31.690 --> 41:35.330 And we have a four-quadrant amplifier. 41:36.170 --> 41:42.850 This four-quadrant amplifier actually allows us to produce any 41:42.850 --> 41:49.230 arbitrary power situation at any arbitrary place in any power grid. 41:49.230 --> 41:54.910 So, it can be a very bad situation, having phase shifts, having extra 41:54.910 --> 41:59.570 or noise, bad noise on the power signal, which means that the power 41:59.570 --> 42:01.130 signal is of a very bad quality. 42:02.070 --> 42:06.950 And then we could try to improve the situation by adequately 42:06.950 --> 42:09.390 controlling the devices in the house. 42:09.390 --> 42:14.230 For example, using the solar inverters in a reasonable way for 42:14.230 --> 42:15.790 providing energy services. 42:16.670 --> 42:28.110 And in this way, we can use this in so-called hardware in the loop 42:28.110 --> 42:35.350 scenario, where we run a simulation of a network and we generate a 42:35.350 --> 42:38.150 certain or we have a certain profile of a house. 42:38.150 --> 42:44.610 And this situation can be mapped onto this real energy smart home lab. 42:45.510 --> 42:50.890 Then we can respond to that with our way of controlling the house and 42:50.890 --> 42:54.950 see how this house can respond to a certain energy situation. 42:54.950 --> 43:02.930 And this can be taken as a model for other nodes in that simulation. 43:03.170 --> 43:08.350 In that way, we can have a simulation which is based on a model which 43:08.350 --> 43:14.990 can adjust it to the real situation that we have in this house. 43:14.990 --> 43:20.510 So, this is a very important part of realistic simulations of what 43:20.510 --> 43:24.330 would happen in a power grid if we would have more of those houses, 43:24.410 --> 43:24.930 for example. 43:26.510 --> 43:34.690 So, all the devices there in that house are connected to some computer 43:34.690 --> 43:39.130 that is connecting all those devices. 43:39.850 --> 43:45.850 And actually, this is not that of much of interest here. 43:46.390 --> 43:48.010 The interesting points are here. 43:48.830 --> 43:52.470 In that house, we have the intelligent appliances which can 43:52.470 --> 43:55.670 communicate with central control and with each other. 43:55.670 --> 44:02.630 They are intelligent in the way that those devices actually can talk 44:02.630 --> 44:03.890 to the environment. 44:04.330 --> 44:10.950 These are so-called meter-at-home devices which can communicate their 44:10.950 --> 44:16.210 current status, their current program and so on to some external 44:16.210 --> 44:16.770 device. 44:17.390 --> 44:21.230 So, they know their current state and they can respond to control. 44:21.370 --> 44:22.650 They can be remotely controlled. 44:23.750 --> 44:28.230 We have an electric vehicle which can be used as a mobile storage with 44:28.230 --> 44:29.330 bidirectional utilization. 44:30.090 --> 44:32.950 The car cannot only drive forwards and backwards. 44:33.350 --> 44:35.550 It can also feed in power and get power. 44:36.590 --> 44:37.110 Bidirectional. 44:38.030 --> 44:41.370 Not just with respect to mobility, but also with respect to power. 44:41.370 --> 44:45.070 And it certainly is a very large consumer or supplier. 44:46.150 --> 44:49.450 And then we have the decentralized power generation from the rooftop 44:49.450 --> 44:51.130 and the combined heat and power plant. 44:51.230 --> 44:54.790 We have the phase change materials and the ceiling that I mentioned 44:54.790 --> 44:55.150 already. 44:55.230 --> 44:56.650 We have the four-quadrant amplifier. 44:57.210 --> 45:02.070 Actually, we also get there a battery simulator which actually can 45:02.070 --> 45:06.370 simulate any battery characteristics and then we can run experiments 45:06.370 --> 45:08.270 with the use of batteries in the house. 45:09.530 --> 45:15.450 So, what we are interested in is the interrelationship between humans, 45:15.630 --> 45:19.790 the residents there, the house, the devices and see how we can 45:19.790 --> 45:23.070 actually make best use of that situation. 45:23.070 --> 45:28.030 And in particular, we would like to discover and exploit degrees of 45:28.030 --> 45:30.570 freedom for energy control in that house. 45:30.770 --> 45:35.150 So, everything that we design there has to support that objective. 45:36.790 --> 45:43.430 We had electric vehicles, like here this Meriva Meragio from Opel. 45:45.030 --> 45:53.210 Meanwhile, it's sitting in a museum of Opel as the first car that was 45:53.210 --> 46:00.950 allowed to drive on the normal streets, having the possibility to 46:00.950 --> 46:04.970 actually refeed power into the grid. 46:05.910 --> 46:11.270 We were supposed to get also an A-Class from Daimler, but that never 46:11.270 --> 46:11.630 happened. 46:11.930 --> 46:15.730 Daimler was not capable to provide us with an electric vehicle. 46:17.030 --> 46:20.390 In particular, they were not capable of providing us with a 46:20.390 --> 46:21.690 bidirectional power connection. 46:22.070 --> 46:24.150 It was only possible together with Opel. 46:25.230 --> 46:30.010 We did get smarts there, but only bidirectional power connection. 46:37.570 --> 46:38.910 They didn't want it. 46:41.110 --> 46:46.670 For reasons that we just heard about before, that they are not that 46:46.670 --> 46:50.970 fond of actually having feedback of power. 46:51.470 --> 46:53.750 So, it was not of high priority for them. 46:53.870 --> 46:54.950 They just didn't find the time. 46:56.070 --> 46:59.410 They certainly would have been capable to do it. 47:00.530 --> 47:01.050 Definitely. 47:01.730 --> 47:03.470 Yes, but they were not interested in that. 47:04.090 --> 47:05.150 So, they didn't do it. 47:07.570 --> 47:12.990 Then, our approach certainly is to have demand-side load management. 47:13.290 --> 47:16.950 Definitely, the appliances can be rescheduled. 47:17.070 --> 47:19.430 We have a 24-hour tariff. 47:19.610 --> 47:23.610 So, we have these time-of-use tariffs. 47:26.490 --> 47:32.390 And certainly, the challenge is that we get an increasing imbalance of 47:32.390 --> 47:33.950 generator and consumer load in the grid. 47:34.070 --> 47:34.850 We know all that. 47:36.990 --> 47:42.610 Then, how can we actually use all those devices inside such a house? 47:42.610 --> 47:46.870 For that, we have to find out what are the capabilities of the 47:46.870 --> 47:47.310 devices. 47:48.030 --> 47:50.270 We have to classify them in some way. 47:51.190 --> 47:57.690 And so, we say that there are some devices which cannot be rescheduled 47:57.690 --> 47:58.190 actually. 47:58.190 --> 48:04.830 They have a very low degree of freedom, like a stove is used at one 48:04.830 --> 48:09.150 specific time during the day and you don't want to shift that time. 48:11.030 --> 48:15.410 So, there are some appliances where rescheduling does not make sense. 48:15.410 --> 48:21.610 There are others, like a deep freezer or electric heating, where you 48:21.610 --> 48:26.410 can shift the use of energy bidirectionally on the timeline. 48:26.770 --> 48:31.450 So, if you have a certain power profile here, you can shift that in 48:31.450 --> 48:32.370 both directions. 48:33.490 --> 48:38.370 Start earlier or stop later. 48:40.350 --> 48:43.970 Some can only be moved backward on the timeline. 48:45.130 --> 48:51.030 That means you can start earlier heating water, but you should not 48:51.030 --> 48:54.890 start later because then the warm water is not available. 48:55.950 --> 48:57.730 So, there are certain deadlines for that. 48:58.430 --> 49:01.510 Or you can move it forward on the timeline. 49:02.110 --> 49:07.450 A washing machine or a dishwasher, like if a dishwasher is at a 49:07.450 --> 49:13.490 certain time T, you could start a dishwasher. 49:13.770 --> 49:17.130 You cannot go to an earlier time because then it is not loaded. 49:17.630 --> 49:21.230 You have to wait for the dishwasher to be loaded and then you can 49:21.230 --> 49:24.530 postpone the start of the dish washing process. 49:24.810 --> 49:28.770 So, you can move it forward on the timeline, but not backwards. 49:29.950 --> 49:32.850 So, there are different constraints to these different devices. 49:34.210 --> 49:39.710 This is just classifying those depliances again into some that are 49:39.710 --> 49:43.890 controllable, others that are only observable and not controllable. 49:44.010 --> 49:46.930 Certainly, those that are controllable are also observable. 49:47.590 --> 49:53.530 Here we have some providing a permanent service, for example, a deep 49:53.530 --> 49:54.550 freezer or a fridge. 49:54.870 --> 50:01.370 They are providing continuous service, whereas we have a timed service 50:01.370 --> 50:04.530 for washing machine, dishwasher and things like that. 50:04.530 --> 50:07.950 You have to know the characteristics of the devices in order to know 50:07.950 --> 50:10.050 how to reschedule them. 50:10.910 --> 50:15.270 And then you have some which are only observable, but you can predict 50:15.270 --> 50:21.270 at which time they will be used, like the stove or a heating plate or 50:21.270 --> 50:22.950 the TV set. 50:23.690 --> 50:26.010 You can also predict when it will be used. 50:28.170 --> 50:29.150 Unpredictable, it says. 50:29.530 --> 50:32.410 Multimedia, TV set, I would say it is more or less predictable. 50:33.430 --> 50:36.430 Lighting, small appliances, there are some that are unpredictable, 50:36.670 --> 50:41.930 that are just switched on when somebody is there, but maybe very 50:41.930 --> 50:42.330 random. 50:43.930 --> 50:47.770 You have to know all those properties in order to be able to use them 50:47.770 --> 50:48.550 for load shifting. 50:49.570 --> 50:54.450 And then you have your time of the day. 50:54.610 --> 51:02.690 You have 24-hour signals, so that's, for example, time of use, tariff 51:02.690 --> 51:05.130 or critical peak pricing or whatever. 51:05.750 --> 51:09.490 Normally it's just the time of use, but can be different types of 51:09.490 --> 51:10.330 tariffs. 51:12.970 --> 51:15.050 Indicated here with this blue line. 51:16.050 --> 51:20.350 Here you have a spread between 10 cents and 40 cents. 51:21.810 --> 51:24.470 So here you have five different levels. 51:25.150 --> 51:26.910 That was the situation in Meridio. 51:27.870 --> 51:32.830 And then you certainly would like to consume energy just during the 51:32.830 --> 51:33.870 lowest rated periods. 51:35.110 --> 51:38.170 This leads to avalanche effects, as we know, but if people do that 51:38.170 --> 51:44.030 manually, they are not behaving exactly the same, so this is working 51:44.030 --> 51:44.830 to some extent. 51:46.550 --> 51:52.670 So now we have a certain schedule, so you could also do that more or 51:52.670 --> 51:53.230 less automatically. 51:53.550 --> 51:57.610 You know, for example, there are certain typical schedules, you have 51:57.610 --> 52:04.970 observed the typical behavior of a certain family, and then you can 52:04.970 --> 52:08.910 predict the schedule that you will have for the next day. 52:09.830 --> 52:14.710 And you have here in yellow now the price signal, and now you could 52:14.710 --> 52:17.550 modify the use of those devices. 52:18.230 --> 52:22.410 So you would like to reschedule the appliances there. 52:23.150 --> 52:30.150 You would like to do that by using the information on the price 52:30.150 --> 52:31.130 signal. 52:33.050 --> 52:37.930 And you would like to find the best starting time concerning this time 52:37.930 --> 52:38.850 signal. 52:42.770 --> 52:45.490 So what kind of signals actually do we have? 52:46.470 --> 52:48.530 As I said, here just the time signal. 52:49.370 --> 52:50.830 Maybe that we have more. 52:51.010 --> 52:55.350 Maybe sometimes we have not just the price, but also some power 52:55.350 --> 52:55.910 constraints. 52:56.850 --> 53:00.630 You have a natural power constraint signal if you would like to 53:00.630 --> 53:05.190 optimize the usage of the power from your rooftop, because then you 53:05.190 --> 53:09.190 don't want to use more power than what is available from your rooftop. 53:10.150 --> 53:14.490 But you can also get power limitation signals from the distribution 53:14.490 --> 53:15.530 system operator. 53:16.830 --> 53:21.370 This is different types of signals, and then the objectives that you 53:21.370 --> 53:26.450 have are to minimize costs, to minimize current power limit 53:26.450 --> 53:26.970 violations. 53:27.170 --> 53:30.710 Maybe that you have to pay for power limit violations, then it is 53:30.710 --> 53:32.070 still minimizing costs. 53:33.010 --> 53:39.630 Maybe that you have sharp constraints, or hard constraints, that 53:39.630 --> 53:44.830 something bad happens if you violate those power limits. 53:45.710 --> 53:49.430 And then you might also have these short term signals, for example, 53:49.470 --> 53:52.110 from this pool of devices that we had. 53:52.110 --> 53:57.130 And we have power factors, we have over and under voltage, we have all 53:57.130 --> 54:01.470 kinds of interesting information that we can respond to. 54:02.390 --> 54:07.770 And we would like to adapt to given values that are shown there for 54:07.770 --> 54:12.210 the energy system and contribute to stabilizing the grid. 54:13.230 --> 54:19.730 Now, what could happen if we make a simple example of demand 54:19.730 --> 54:20.290 management? 54:20.550 --> 54:26.970 If you have such an original predicted schedule, and now you have the 54:26.970 --> 54:31.590 price signal, and then you would like to do certain things, the bread 54:31.590 --> 54:37.210 machine that is starting there should start earlier, because the bread 54:37.210 --> 54:42.190 machine can start earlier in time if it's loaded in the evening. 54:42.210 --> 54:46.010 It should produce the bread for the morning, for the breakfast. 54:46.330 --> 54:49.130 That could be done at some earlier time. 54:49.990 --> 54:55.850 Then maybe you have to start the dishwasher after lunch. 54:56.150 --> 55:00.090 That can be postponed to a time where energy is cheaper. 55:01.430 --> 55:04.970 And so you put it here to a time where you have the low price. 55:06.230 --> 55:11.970 And then you have again here maybe a dishwasher or washing machine, 55:12.230 --> 55:14.970 again they are postponed to a different time when it's cheaper. 55:15.230 --> 55:18.350 This is a simple way of actually rescheduling. 55:18.930 --> 55:23.810 Here we only look at the price for optimizing. 55:24.640 --> 55:28.930 And then we have the stove usage, which cannot be moved, but it is a 55:28.930 --> 55:30.570 high price period. 55:31.030 --> 55:34.890 You use power from the electric vehicle, and then maybe you can 55:34.890 --> 55:39.450 recharge that at night times where you have cheap power again. 55:40.090 --> 55:44.270 So this is the use of a battery, maybe a stationary battery or a 55:44.270 --> 55:45.010 mobile battery. 55:45.830 --> 55:51.770 So this is just showing what you could do during a day, postpone or 55:51.770 --> 55:56.570 schedule the devices at reasonable times of the day. 55:57.750 --> 56:03.050 This is done in the house, in that energy smart home lab on campus. 56:03.630 --> 56:05.990 You see that it is a nice home inside. 56:07.750 --> 56:11.330 Has anybody of you been in that already? 56:11.670 --> 56:13.890 Because sometimes there are visiting periods. 56:14.730 --> 56:20.590 Maybe we can arrange a visit at some later time in this course. 56:20.590 --> 56:24.530 And then inside that house we have this energy management panel. 56:25.010 --> 56:27.030 What does the energy management panel do? 56:27.850 --> 56:32.090 It is supposed to provide information on energy consumption, make the 56:32.090 --> 56:35.330 resident aware of the current energy situation. 56:36.210 --> 56:41.430 The other point is to discover and specify degrees of freedom for 56:41.430 --> 56:42.330 energy consumption. 56:43.950 --> 56:51.290 The energy management system in the house needs to know what the 56:51.290 --> 56:55.730 preferences of the user are for, for example, using the dishwasher or 56:55.730 --> 56:56.930 using the electric vehicle. 56:56.930 --> 57:01.690 And so you need appropriate interfaces in such a system. 57:02.950 --> 57:04.610 So what kind of things do you see? 57:04.670 --> 57:07.870 By the way, the energy management panel is just a web application. 57:08.310 --> 57:12.650 That means it can be displayed on any device that has the capability 57:12.650 --> 57:16.390 to look at HTML pages. 57:17.230 --> 57:22.430 And so you can have it on some device, preferably on a touch screen, 57:23.270 --> 57:25.070 but you can also have it on your smartphone. 57:25.770 --> 57:33.050 And so this here is showing the current energy situation. 57:33.950 --> 57:40.550 I will actually now try to go directly... Oops, I forgot to change 57:40.550 --> 57:41.310 something there. 57:42.490 --> 57:44.590 I noticed this error. 57:46.190 --> 57:51.450 This is just a simple error. 57:52.670 --> 57:59.090 So there was an HTTPS in front, or should be an HTTPS in front. 57:59.090 --> 58:02.310 So this is the current situation in the house. 58:04.110 --> 58:08.590 In the Energy Smart Home Lab, you see a certain price for energy, 58:08.710 --> 58:10.210 which is not the real price. 58:10.990 --> 58:15.950 It is a locally generated price, but here you see the price curve for 58:15.950 --> 58:18.190 the next 24 hours. 58:18.190 --> 58:24.590 You also see in red the power constraint, which we also generate 58:24.590 --> 58:26.450 there. 58:26.630 --> 58:29.670 It is just locally. 58:30.990 --> 58:36.610 And here you see the consumption for the last 24 hours. 58:37.330 --> 58:41.450 So you see that something had been going on at noontime. 58:42.570 --> 58:46.530 Probably some people from my group had their lunch there. 58:47.370 --> 58:50.310 So I can immediately see if they have been going there or not. 58:52.450 --> 58:55.790 There you see the current power consumption. 58:58.050 --> 59:06.150 And so what I can also see is I can click on that consumption history 59:06.150 --> 59:08.650 and look at all the different devices. 59:08.650 --> 59:15.950 For example, I can set any date. 59:16.030 --> 59:21.710 I could say I would like to start from the 1st of May and see, for 59:21.710 --> 59:30.810 example, how often the washing machine has been used. 59:31.970 --> 59:33.430 I get nothing. 59:34.690 --> 59:36.070 Oh, there it is. 59:36.230 --> 59:36.470 You see? 59:37.570 --> 59:41.810 So quite often the dishwasher has been used. 59:42.670 --> 59:49.210 So you can get, in a flexible way, information on the consumption 59:49.210 --> 59:52.090 history of all the devices in a house. 59:52.990 --> 59:59.010 You can also click in this area and get an overview of the current 59:59.010 --> 01:00:00.450 power flow in the house. 01:00:01.510 --> 01:00:08.530 So you see that we get 645 watt from the system. 01:00:09.230 --> 01:00:10.670 Something is missing here. 01:00:11.170 --> 01:00:12.630 There is an empty space. 01:00:12.630 --> 01:00:16.210 The photovoltaic panels used to be there, but they are not on the 01:00:16.210 --> 01:00:18.850 rooftop at the moment, so we don't have power flow from there. 01:00:20.370 --> 01:00:25.370 But in this way you get information on how the power currently is 01:00:25.370 --> 01:00:25.630 flowing. 01:00:25.710 --> 01:00:30.110 You see that the air conditioning actually is working at the moment. 01:00:31.550 --> 01:00:36.490 The ceiling is cooled by pumping cool water actually through that. 01:00:37.790 --> 01:00:43.310 And I can click, for example, here on one of those icons and I get 01:00:43.310 --> 01:00:44.650 information on that device. 01:00:44.650 --> 01:00:53.310 I can also look at the layout of the house there. 01:00:56.950 --> 01:01:00.250 The car used to be there. 01:01:00.410 --> 01:01:04.570 It's no longer there because our project ended and Opel took it away 01:01:04.570 --> 01:01:05.910 and put it in a museum. 01:01:06.810 --> 01:01:09.150 But we will get other cars there in the future. 01:01:12.230 --> 01:01:12.830 Yes? 01:01:16.680 --> 01:01:18.260 Yes, definitely. 01:01:18.840 --> 01:01:21.400 If we don't use it locally, we just feed it back. 01:01:21.700 --> 01:01:30.100 We have not actually registered that with the utility company, so we 01:01:30.100 --> 01:01:31.580 don't get paid for that. 01:01:32.360 --> 01:01:32.980 We just do it. 01:01:34.700 --> 01:01:37.960 So this is a layout and we can click on any area there. 01:01:38.060 --> 01:01:44.500 We could actually also, if I go on the light symbol here, then I could 01:01:44.500 --> 01:01:45.480 do the following. 01:01:45.820 --> 01:01:48.920 I could say, OK, I just switch on the light. 01:01:49.140 --> 01:01:51.740 Now, something like that should not be done, really. 01:01:51.820 --> 01:01:58.100 I hope that nobody is in the house at the moment, so I really should 01:01:58.100 --> 01:01:58.900 switch it off again. 01:01:59.600 --> 01:02:04.040 So this is a feature that we can actually have remote control of the 01:02:04.040 --> 01:02:04.440 house. 01:02:05.140 --> 01:02:09.160 It means that on your smartphone you can just click on it and you can 01:02:09.160 --> 01:02:11.220 switch on or off the light. 01:02:11.740 --> 01:02:16.420 This button here, this off button, was an important additional 01:02:16.420 --> 01:02:16.820 feature. 01:02:16.820 --> 01:02:20.380 The first residents of that home, so we have living periods there of 01:02:20.380 --> 01:02:25.300 test persons, and the first residents said they would like to have the 01:02:25.300 --> 01:02:30.720 possibility to switch off everything to make sure that if they notice, 01:02:30.860 --> 01:02:36.680 oh, they left the light on, then they would like to be able to switch 01:02:36.680 --> 01:02:37.160 it off. 01:02:37.480 --> 01:02:39.460 So we provided them with that capability. 01:02:40.060 --> 01:02:42.800 So this is just a little bit of home automation. 01:02:43.600 --> 01:02:47.700 Then, in this layout, we can... there are also some other information. 01:02:48.380 --> 01:02:50.620 These are all the different devices. 01:02:51.560 --> 01:02:59.920 This here is giving temperature information, and then we can click, 01:03:00.040 --> 01:03:05.260 for example, on the kitchen area, and we get information on the 01:03:05.260 --> 01:03:09.240 different devices in that kitchen area. 01:03:09.240 --> 01:03:12.500 So we have a stove, we have a washing machine, we have a dishwasher, 01:03:12.820 --> 01:03:18.540 we have a tumble dryer, we have a coffee machine, and so on. 01:03:18.540 --> 01:03:25.240 And then I can go on any of those devices, for example, here on the 01:03:25.240 --> 01:03:31.800 washing machine, and then what you see is the interface for remote 01:03:31.800 --> 01:03:33.200 control of that device. 01:03:33.400 --> 01:03:34.740 You see the current situation. 01:03:34.740 --> 01:03:37.100 So currently it's not active. 01:03:37.340 --> 01:03:39.180 There's no laundry in there. 01:03:39.820 --> 01:03:47.260 But if you put in laundry, then you would get... and if you open that 01:03:47.260 --> 01:03:52.640 menu, you would get an additional button, which would say start. 01:03:52.640 --> 01:03:55.920 Now, you don't have to use that, but you can use that. 01:03:56.020 --> 01:03:57.360 I'll come back to that in a moment. 01:03:57.680 --> 01:04:04.880 What you're supposed to do is you just set the time period that you 01:04:04.880 --> 01:04:08.360 would like to use for that washing process. 01:04:09.240 --> 01:04:13.580 So you say the next four and a half hours, for example, are sufficient 01:04:13.580 --> 01:04:14.120 for that. 01:04:14.740 --> 01:04:20.360 So four and a half hours from now is... it says 9-11. 01:04:20.560 --> 01:04:24.640 It's not exactly, but we could modify that. 01:04:24.640 --> 01:04:32.660 I can reduce that easily and in that way set the deadline for actually 01:04:32.660 --> 01:04:35.980 having that washing program finished. 01:04:37.500 --> 01:04:43.660 And then, if that machine has been initialized, the system would come 01:04:43.660 --> 01:04:49.620 back with information on the starting time that will be selected for 01:04:49.620 --> 01:04:50.720 an optimized schedule. 01:04:50.720 --> 01:04:57.900 So this loading of the machine leads to an optimization of the... or 01:04:57.900 --> 01:05:03.660 setting that degree of freedom here leads to a request for the energy 01:05:03.660 --> 01:05:08.700 management system and will come back with a desired, optimized 01:05:08.700 --> 01:05:10.660 starting time. 01:05:11.120 --> 01:05:13.940 And if you don't like that, you can switch on the start button and 01:05:13.940 --> 01:05:14.700 start it immediately. 01:05:15.570 --> 01:05:19.760 So the philosophy behind that is to give a lot of information to the 01:05:19.760 --> 01:05:25.700 user, let it specify the preferences, but don't restrict the user in 01:05:25.700 --> 01:05:30.280 his or her freedom to use the device. 01:05:30.680 --> 01:05:33.220 You can just switch on start and then it starts. 01:05:34.180 --> 01:05:40.920 This start button actually was not there originally, but was put in on 01:05:40.920 --> 01:05:45.920 a request of the residents who said they don't want to be restricted 01:05:45.920 --> 01:05:49.740 just to use this schedule or this optimizer here. 01:05:50.240 --> 01:05:57.740 They would like to be able to start it by just clicking on some button 01:05:57.740 --> 01:05:58.060 there. 01:05:58.060 --> 01:06:02.980 So we put it in and that button has never been used, but the people 01:06:02.980 --> 01:06:06.740 were very satisfied that now they had the option to actually use it. 01:06:07.460 --> 01:06:11.400 And this is an important effect that if you do something like that, 01:06:11.480 --> 01:06:15.880 you have to look at the psychology of using such an interface. 01:06:15.880 --> 01:06:22.020 You should never give the impression that the user is restricted in 01:06:22.020 --> 01:06:24.580 his freedom to use certain devices. 01:06:26.200 --> 01:06:31.100 If the user likes to, he can easily specify the degrees of freedom on 01:06:31.100 --> 01:06:34.360 the fly and then we have the information. 01:06:34.360 --> 01:06:38.780 Okay, so this is a typical schedule there. 01:06:39.260 --> 01:06:46.940 Then we have here also the information on the larger systems there. 01:06:47.960 --> 01:06:52.860 And for example, for the car, there we have a schedule or some menu 01:06:52.860 --> 01:07:01.100 where you get information on the current range, like the state of 01:07:01.100 --> 01:07:05.680 charge of the battery, or translate it into the range that is 01:07:05.680 --> 01:07:06.080 available. 01:07:06.800 --> 01:07:12.100 You can specify the next starting time. 01:07:13.440 --> 01:07:19.120 So you could say, okay, I would like to start at some later time. 01:07:20.160 --> 01:07:27.340 And if I go sufficiently far here, then it should actually switch also 01:07:27.340 --> 01:07:27.800 the date. 01:07:27.800 --> 01:07:30.300 I never used that. 01:07:30.400 --> 01:07:30.940 Yes, you see? 01:07:31.460 --> 01:07:33.100 Then you can set just any time. 01:07:33.200 --> 01:07:36.480 You could also certainly just say, okay, I would like to go to the 01:07:36.480 --> 01:07:36.900 next day. 01:07:37.560 --> 01:07:42.920 So this is a simple way of actually setting the starting time or the 01:07:42.920 --> 01:07:46.000 next driving time for that vehicle. 01:07:46.560 --> 01:07:53.120 And you can specify the minimum range in some intervals. 01:07:53.120 --> 01:07:57.700 So that's providing you with the capability to say, I leave the car 01:07:57.700 --> 01:07:58.320 here now. 01:07:58.640 --> 01:07:59.980 You can use the battery. 01:08:00.260 --> 01:08:06.460 I would like to have a fully charged battery at just after midnight 01:08:06.460 --> 01:08:08.220 the next day. 01:08:08.460 --> 01:08:12.460 And all the time I would like to have a minimum state of charge such 01:08:12.460 --> 01:08:14.760 that I can drive for 27 kilometers. 01:08:15.840 --> 01:08:19.840 And so in this way, the user specifies all his requirements. 01:08:20.600 --> 01:08:25.760 And the energy management system has all the information on the 01:08:25.760 --> 01:08:30.020 flexibility of that battery and can utilize that afterwards. 01:08:32.040 --> 01:08:34.720 So this is also an important thing. 01:08:35.420 --> 01:08:38.240 I don't go into the other devices here. 01:08:38.300 --> 01:08:40.960 That was all that is of interest in the moment. 01:08:41.440 --> 01:08:42.780 We can get back to these. 01:08:42.860 --> 01:08:46.260 So these are the slides showing exactly the things that I just showed 01:08:46.260 --> 01:08:46.420 you. 01:08:46.540 --> 01:08:53.420 So this is another menu showing the history of the power generation 01:08:53.420 --> 01:08:54.880 from photovoltaic systems. 01:08:54.880 --> 01:08:56.740 You just saw that. 01:08:57.140 --> 01:08:57.920 Here it's in English. 01:08:58.860 --> 01:09:06.580 Here's again this menu where you have also this extra button that you 01:09:06.580 --> 01:09:10.600 can click on in order to start a system immediately. 01:09:11.780 --> 01:09:17.960 So this is an important part to get the preferences of the user and to 01:09:17.960 --> 01:09:22.460 inform the user about what's happening in the system or in the house. 01:09:24.060 --> 01:09:27.440 Again, here the air conditioning with this phase change material, 01:09:28.160 --> 01:09:35.340 small packets of some kind of salt which crystallizes or melts and the 01:09:35.340 --> 01:09:40.000 melting is producing or is using energy and this is a cooling effect. 01:09:41.000 --> 01:09:47.400 You have seen these things and in the technology room we have all 01:09:47.400 --> 01:09:50.600 kinds of devices which control the individual elements. 01:09:50.740 --> 01:09:56.580 So in that house every location that is of interest is connected with 01:09:56.580 --> 01:10:03.440 respect to the energy consumption and also some information about 01:10:03.440 --> 01:10:06.620 what's happening there so we can control there many different 01:10:06.620 --> 01:10:08.280 locations inside the house. 01:10:08.860 --> 01:10:15.100 But it is a research environment so we could afford to actually have 01:10:15.100 --> 01:10:17.760 such an enormous amount of wiring there. 01:10:18.760 --> 01:10:25.080 In the background the energy management system is actually called 01:10:25.080 --> 01:10:27.500 organic smart home meanwhile. 01:10:28.480 --> 01:10:33.820 Organic smart home because we use an architecture that has been 01:10:33.820 --> 01:10:39.420 designed in our research program on organic computing some time ago. 01:10:40.420 --> 01:10:45.540 Organic computing is on self-organizing systems and we noticed that 01:10:45.540 --> 01:10:50.780 the energy system is full of self-organization and all kinds of 01:10:50.780 --> 01:10:56.680 decentralized devices which should have some local intelligence and 01:10:56.680 --> 01:11:06.140 the architecture of organic computing was perfect for the energy 01:11:06.140 --> 01:11:08.020 application area. 01:11:08.020 --> 01:11:09.460 So what do you see here? 01:11:09.940 --> 01:11:15.740 You see devices, fridge or washing machine. 01:11:16.200 --> 01:11:17.640 They have certain drivers. 01:11:17.800 --> 01:11:22.300 They can communicate with their environment using the Miele at home 01:11:22.300 --> 01:11:24.760 protocol for example but could also be other ones. 01:11:24.760 --> 01:11:29.880 You have a hardware abstraction layer that is not the computer here, 01:11:30.080 --> 01:11:30.980 the hell computer. 01:11:31.540 --> 01:11:33.840 Who of you knows what hell computer is? 01:11:35.540 --> 01:11:38.780 Has no one of you ever heard about hell, the computer? 01:11:38.780 --> 01:11:48.040 Has anyone of you read the book about hitchhiking in the galaxies? 01:11:49.360 --> 01:11:49.540 No? 01:11:50.400 --> 01:11:51.120 Nice books. 01:11:52.980 --> 01:11:55.960 Some science fiction things. 01:11:55.960 --> 01:12:00.900 So if you take... hell was the name of the computer in that science 01:12:00.900 --> 01:12:02.040 fiction novel. 01:12:02.660 --> 01:12:10.440 If you just modify the letters by one or just take the next one, you 01:12:10.440 --> 01:12:13.100 can easily find out what hell then stands for. 01:12:14.160 --> 01:12:18.640 If you go to the next letters, then you have a different name, which 01:12:18.640 --> 01:12:19.400 is quite well known. 01:12:20.720 --> 01:12:23.560 So we have here this hardware abstraction layer, that's the 01:12:23.560 --> 01:12:25.100 abbreviation which is hell. 01:12:25.740 --> 01:12:29.820 The hardware abstraction layer provides information, essential 01:12:29.820 --> 01:12:33.000 information, for example about a washing machine or about a freezer. 01:12:33.000 --> 01:12:39.900 And then this is connected to some observer unit and controller unit, 01:12:40.080 --> 01:12:44.700 so it is observed what's happening in that washing machine and you can 01:12:44.700 --> 01:12:46.980 perform certain control actions. 01:12:47.780 --> 01:12:53.740 And this information is forwarded to some centralized observer, which 01:12:53.740 --> 01:12:58.260 will notice what kind of things are going on in the house and from 01:12:58.260 --> 01:13:02.820 that it will predict what will be happening for the next day, for 01:13:02.820 --> 01:13:03.220 example. 01:13:03.220 --> 01:13:10.840 And this information will be used by a controller, which will optimize 01:13:10.840 --> 01:13:12.280 the schedule for the day. 01:13:12.400 --> 01:13:18.260 I said the energy management system provides on this menu the time 01:13:18.260 --> 01:13:19.780 when this device will be started. 01:13:20.840 --> 01:13:24.600 This is done by that controller after an optimization process. 01:13:24.600 --> 01:13:29.640 And then we have communication with the human user, with the residents 01:13:29.640 --> 01:13:34.820 of the house, and also with external elements, external entities, 01:13:35.340 --> 01:13:42.000 providing price signals, power constraint signals, and weather signals 01:13:42.000 --> 01:13:45.600 and things like that, like weather information, whatever might be 01:13:45.600 --> 01:13:46.400 necessary there. 01:13:46.400 --> 01:13:51.960 So this architecture is essential for observing individual devices, 01:13:52.360 --> 01:13:56.100 observing and predicting what's happening in the house, learning and 01:13:56.100 --> 01:14:01.800 controlling what should be done in the house and how one should 01:14:01.800 --> 01:14:05.160 influence what's happening in the house, and you have to be able to 01:14:05.160 --> 01:14:05.640 communicate. 01:14:05.640 --> 01:14:12.080 So that's this architecture, which is doing all the management in the 01:14:12.080 --> 01:14:12.360 house. 01:14:12.940 --> 01:14:17.300 And this is software, which is meanwhile available as an open source 01:14:17.300 --> 01:14:22.020 software, has a SourceForge license and can be used by anybody. 01:14:22.640 --> 01:14:26.960 And it's running at least in the Energy Smart Home Lab on campus. 01:14:27.340 --> 01:14:32.200 It's also running in the FZI House of Living Labs in an extended 01:14:32.200 --> 01:14:32.600 version. 01:14:32.600 --> 01:14:39.120 And I think it's also used at some other place, but this is running 01:14:39.120 --> 01:14:39.620 software. 01:14:40.280 --> 01:14:43.120 So it's not just a concept, but this actually is running. 01:14:43.680 --> 01:14:48.600 And in the House of Living Labs, we are responsible there for the 01:14:48.600 --> 01:14:52.820 energy management for electricity and heat. 01:14:53.760 --> 01:14:55.460 And this is all done with these systems. 01:14:56.520 --> 01:15:02.120 The background for that is, as I mentioned already, the name Organic 01:15:02.120 --> 01:15:04.280 Smart Home comes from organic computing. 01:15:05.240 --> 01:15:08.560 And so I would like to briefly tell you something about this 01:15:08.560 --> 01:15:12.700 background, which influenced the way we designed this system. 01:15:12.700 --> 01:15:15.300 So organic computing. 01:15:16.220 --> 01:15:23.680 This actually is the cover of a book, which contains all the results 01:15:23.680 --> 01:15:25.180 of that research program. 01:15:26.080 --> 01:15:28.780 You can download it from SpringerLink. 01:15:29.060 --> 01:15:31.320 It's available to you free of charge. 01:15:31.700 --> 01:15:38.940 If you enter it with your KIT account, you can download it and you get 01:15:38.940 --> 01:15:42.320 all kinds of information on the technology that we have designed 01:15:42.320 --> 01:15:42.580 there. 01:15:43.520 --> 01:15:47.200 So let me briefly tell you something about organic computing. 01:15:48.220 --> 01:15:52.020 This is actually something which happened at the beginning of this 01:15:52.020 --> 01:15:53.400 century in 2002. 01:15:54.140 --> 01:15:58.040 The people in the computer engineering sections of the Gesellschaft 01:15:58.040 --> 01:16:06.620 der Informatik and the Information Technology Society of VDE. 01:16:07.680 --> 01:16:14.220 We had several workshops on challenges for research. 01:16:16.580 --> 01:16:17.340 And thank you. 01:16:18.820 --> 01:16:25.920 And we looked at what are the challenges for the next 10 or 15 years. 01:16:26.720 --> 01:16:31.180 So we noticed, okay, we are moving towards ubiquitous network 01:16:31.180 --> 01:16:31.700 computer. 01:16:32.640 --> 01:16:37.420 And these complex, ubiquitous systems need definitely new concepts for 01:16:37.420 --> 01:16:39.640 communicating, for controlling those devices. 01:16:40.700 --> 01:16:43.720 We don't want to be caught in such a network. 01:16:43.820 --> 01:16:46.280 We would like to be able to control it in a reasonable way. 01:16:47.240 --> 01:16:51.280 And so we said that future computer systems have to be designed with 01:16:51.280 --> 01:16:54.760 respect to human needs, whatever that means. 01:16:54.980 --> 01:17:00.580 And it's completely open how those human needs are actually specified 01:17:00.580 --> 01:17:05.540 or how the systems get aware of human needs. 01:17:06.520 --> 01:17:10.580 Those systems should be robust, adaptive and flexible. 01:17:11.680 --> 01:17:18.300 If you have many devices and they break down very often, you are busy 01:17:18.300 --> 01:17:19.620 repairing those devices. 01:17:19.880 --> 01:17:20.880 That's not feasible. 01:17:21.400 --> 01:17:25.360 They should be more or less self-maintaining. 01:17:25.360 --> 01:17:29.060 They should adapt to what you want to use them for. 01:17:29.740 --> 01:17:32.520 And so they should be quite flexible. 01:17:33.380 --> 01:17:35.260 They should also be self-organized. 01:17:35.440 --> 01:17:41.220 That means you don't want to reprogram them every time you would like 01:17:41.220 --> 01:17:42.220 to add a new feature. 01:17:42.760 --> 01:17:46.720 Those devices should be able to reorganize themselves. 01:17:47.380 --> 01:17:50.860 Noticing new requirements, they should be able to transform new 01:17:50.860 --> 01:17:55.760 requirements into new programs that are generated by themselves in 01:17:55.760 --> 01:17:56.200 some way. 01:17:57.540 --> 01:18:03.220 And if they are able to be self-organized, they should still be 01:18:03.220 --> 01:18:03.680 trustworthy. 01:18:04.420 --> 01:18:09.540 Because, you know, if you design an intelligent agent, the typical 01:18:09.540 --> 01:18:13.360 short story is about the intelligent agent that turns against you. 01:18:14.360 --> 01:18:15.920 And so they should be trustworthy. 01:18:15.920 --> 01:18:19.960 It means that they should always do what you expect them to do. 01:18:20.840 --> 01:18:24.760 This is something which we didn't have as a requirement for 01:18:24.760 --> 01:18:29.120 information technology systems so far because we designed them for a 01:18:29.120 --> 01:18:32.860 specific purpose and then they would perform that service. 01:18:32.860 --> 01:18:37.340 Now, if we assume them to be adaptive and flexible and respond to new 01:18:37.340 --> 01:18:40.860 requirements, we must make sure that they don't respond to 01:18:40.860 --> 01:18:44.420 requirements that we don't like or that we don't give them but others 01:18:44.420 --> 01:18:45.840 gave those requirements. 01:18:46.570 --> 01:18:50.800 So this is a difficult topic how you can make sure that systems 01:18:50.800 --> 01:18:52.600 actually stay trustworthy. 01:18:54.080 --> 01:18:57.300 And if you look at those requirements, it's obvious that the behavior 01:18:57.300 --> 01:19:00.600 of those systems should be quite organic. 01:19:01.600 --> 01:19:05.860 An organically behaving system obviously is a system which is 01:19:05.860 --> 01:19:09.340 adaptive, flexible, self-organized and things like that. 01:19:09.940 --> 01:19:12.980 That's why we call them organic computing systems. 01:19:13.940 --> 01:19:20.160 And this was done at the end of 2002 and certainly was influenced by 01:19:20.160 --> 01:19:21.080 other initiatives. 01:19:21.420 --> 01:19:24.320 There were initiatives on ubiquitous computing, autonomic computing, 01:19:24.640 --> 01:19:29.460 pervasive computing, all kinds of different technologies having to do 01:19:29.460 --> 01:19:33.060 with similar scenarios like those. 01:19:34.300 --> 01:19:39.240 So, for organic computing, the challenge was not to say, okay, how can 01:19:39.240 --> 01:19:42.240 we come up with adaptive and self-organizing systems. 01:19:42.380 --> 01:19:46.560 We know that we will be confronted with adaptive and self-organizing 01:19:46.560 --> 01:19:51.880 systems but we have to make sure that they are designed and 01:19:51.880 --> 01:19:54.560 controllable in a reasonable way. 01:19:54.560 --> 01:19:59.800 If we cannot make sure that that's happening, we wouldn't like that 01:19:59.800 --> 01:20:00.300 development. 01:20:01.280 --> 01:20:04.720 And so that was the major challenge and so there are all kinds of 01:20:04.720 --> 01:20:11.020 questions how we can control that, different topics that have to be 01:20:11.020 --> 01:20:17.260 addressed, what kind of control, what kind of policies do we need 01:20:17.260 --> 01:20:22.760 there, how would we specify the behavior, many different topics have 01:20:22.760 --> 01:20:23.280 to be addressed. 01:20:24.560 --> 01:20:30.540 And in Germany, we were lucky to have this priority program running 01:20:30.540 --> 01:20:35.500 for six years where we said we would like to provide a toolbox for 01:20:35.500 --> 01:20:36.860 designing these systems. 01:20:37.420 --> 01:20:41.540 For that, we need knowledge about the principles for setting up 01:20:41.540 --> 01:20:43.000 organic computing systems. 01:20:43.000 --> 01:20:49.840 And in designing those principles, we should be aware of certain 01:20:49.840 --> 01:20:55.520 things that happen in nature because nature has remarkable self 01:20:55.520 --> 01:21:00.520 -organizing systems and so it's interesting to see how those work and 01:21:00.520 --> 01:21:05.300 then maybe that's something which can inspire us in the way we design 01:21:05.300 --> 01:21:06.320 the technical systems. 01:21:07.200 --> 01:21:11.020 And in everything what we do, we should have in mind technical 01:21:11.020 --> 01:21:14.040 applications because that's what is driving us. 01:21:14.500 --> 01:21:18.240 We would like to use or to have an environment where we have technical 01:21:18.240 --> 01:21:22.300 applications which are equipped with many intelligent devices and they 01:21:22.300 --> 01:21:25.080 should have exactly the properties that we mentioned. 01:21:26.480 --> 01:21:33.000 And so this was our program, ran for six years. 01:21:33.700 --> 01:21:35.920 We had around 18 projects. 01:21:36.680 --> 01:21:43.160 It means we had several, I think we had 23 universities which were 01:21:43.160 --> 01:21:51.720 involved in that and quite a few researchers who worked there and came 01:21:51.720 --> 01:21:55.060 up with nice results which were all summarized. 01:21:56.080 --> 01:22:04.060 Now, for self-organizing systems, there have been several approaches 01:22:04.060 --> 01:22:08.320 to have some kind of control systems for that. 01:22:09.040 --> 01:22:13.680 There is the Autonomic Computing Initiative where they actually 01:22:13.680 --> 01:22:17.160 designed something which they call MAPE or MAPE-K. 01:22:17.160 --> 01:22:28.380 Where they say on top of such a device there should be some autonomic 01:22:28.380 --> 01:22:36.500 manager which has the capability to monitor, to analyze, to plan and 01:22:36.500 --> 01:22:40.240 to execute certain actions whenever it's necessary. 01:22:41.180 --> 01:22:43.560 So that is this MAPE cycle. 01:22:43.740 --> 01:22:45.960 Monitoring, analyzing, planning and executing. 01:22:46.880 --> 01:22:50.480 This is an autonomic element which is capable to do something in a 01:22:50.480 --> 01:22:51.440 self -organized way. 01:22:52.460 --> 01:22:57.440 IBM addressed mainly large server architectures, data centers and so 01:22:57.440 --> 01:22:57.600 on. 01:22:57.660 --> 01:23:02.040 They did not address the autonomic computing technical applications. 01:23:03.400 --> 01:23:06.260 Now, we addressed technical applications. 01:23:06.800 --> 01:23:12.220 It's just an arbitrary system on observation and control which can run 01:23:12.220 --> 01:23:14.820 without that control element on top. 01:23:15.260 --> 01:23:20.740 But it should have something or should benefit from having this 01:23:20.740 --> 01:23:22.800 observer -controller architecture on top. 01:23:22.800 --> 01:23:24.480 So this is an observer. 01:23:24.680 --> 01:23:28.080 You could say, well, to monitor and analyze is the same. 01:23:28.720 --> 01:23:31.880 We have a controller similar to planning and executing. 01:23:32.120 --> 01:23:39.680 So very similar but we specified a few more things to a larger extent. 01:23:39.680 --> 01:23:41.620 And we had an additional feature. 01:23:41.840 --> 01:23:49.480 We said that we always need an interface to an external entity to a 01:23:49.480 --> 01:23:56.080 user who would provide goals, objectives for running the system and 01:23:56.080 --> 01:24:00.040 who would also be interested in some status information of that 01:24:00.040 --> 01:24:00.400 system. 01:24:00.500 --> 01:24:05.200 So we need that communication to some higher level entity. 01:24:05.200 --> 01:24:11.700 There certainly is also some communication of the system with its 01:24:11.700 --> 01:24:15.960 environment, but this is a different thing from having an explicit 01:24:15.960 --> 01:24:22.000 external entity providing the objectives for that system. 01:24:24.280 --> 01:24:29.480 Now, we have in this generic architecture that we designed, we said we 01:24:29.480 --> 01:24:32.520 need in the observer a certain number of components. 01:24:32.840 --> 01:24:36.600 Or in a generic setting, it is reasonable to have the following 01:24:36.600 --> 01:24:42.240 components, which in a special setting might not all be necessary and 01:24:42.240 --> 01:24:46.580 have to be implemented in a specific way, but this is just a concept 01:24:46.580 --> 01:24:48.500 telling we need these kinds of systems. 01:24:48.500 --> 01:24:55.460 We need to be able to monitor, to get raw data from the system under 01:24:55.460 --> 01:24:56.520 observation and control. 01:24:57.440 --> 01:25:03.360 We need some kind of pre-processing, taking out some outliers or 01:25:03.360 --> 01:25:08.360 filling in missing data or transforming the data in some way. 01:25:08.360 --> 01:25:13.440 Getting from very high resolution to less high resolution, from second 01:25:13.440 --> 01:25:19.940 resolution to a minute resolution of energy consumption details, for 01:25:19.940 --> 01:25:20.320 example. 01:25:21.140 --> 01:25:25.960 And then we need some kind of data analyzer, which should notice 01:25:25.960 --> 01:25:31.520 certain patterns, specific things so a data analyzer in the energy 01:25:31.520 --> 01:25:36.740 scenario could be able to find out that a washing machine has been 01:25:36.740 --> 01:25:37.180 started. 01:25:37.180 --> 01:25:42.280 It could have information on the typical pattern of a washing machine 01:25:42.280 --> 01:25:45.260 and could notice, oh, there has been a washing machine that has been 01:25:45.260 --> 01:25:45.640 started. 01:25:46.720 --> 01:25:48.140 Or some other things. 01:25:48.800 --> 01:25:53.280 We should also have a predictor, a prediction component, which can 01:25:53.280 --> 01:25:59.180 predict what will be happening in the near future of that system. 01:26:00.120 --> 01:26:03.020 And then, so this is what they should do. 01:26:03.560 --> 01:26:06.960 Monitor, we should also be able to store data over some time. 01:26:08.360 --> 01:26:16.340 And then we should also be able to maybe modify the way we observe the 01:26:16.340 --> 01:26:20.420 system, because it may be that for some time it's sufficient to 01:26:20.420 --> 01:26:24.780 monitor or to look at a system every five minutes. 01:26:24.780 --> 01:26:26.820 And now you see a change. 01:26:27.940 --> 01:26:29.520 You will monitor it every minute. 01:26:29.920 --> 01:26:32.200 You will see more changes. 01:26:32.400 --> 01:26:39.300 You will increase the resolution in order to be able to control it in 01:26:39.300 --> 01:26:40.980 a more fine-grained way. 01:26:41.540 --> 01:26:44.140 So you would modify the model of observation. 01:26:44.340 --> 01:26:49.040 Also, maybe you look at different components depending on what you 01:26:49.040 --> 01:26:49.740 have seen before. 01:26:50.740 --> 01:26:55.500 And then you need also the controller. 01:26:56.220 --> 01:27:00.740 And in the controller you have the mapping, which takes a situation 01:27:00.740 --> 01:27:07.240 that has been observed by the observer, is put into this mapping, and 01:27:07.240 --> 01:27:13.000 this mapping now has a way of mapping a situation onto actions. 01:27:13.780 --> 01:27:18.060 And now it may happen that several actions are activated, and you have 01:27:18.060 --> 01:27:19.800 to find out which one is the best. 01:27:20.540 --> 01:27:24.000 So you need some information on adequacy of an action in a certain 01:27:24.000 --> 01:27:28.620 situation, and there's an action selector which actually makes sure 01:27:28.620 --> 01:27:33.240 that this action can be executed on that system under observational 01:27:33.240 --> 01:27:33.720 control. 01:27:34.680 --> 01:27:41.640 And then you would also store a history of what you have done before, 01:27:42.640 --> 01:27:49.360 and you compare the current situation with what you have done 01:27:49.360 --> 01:27:53.280 previously, and then you can see the impact of what your previous 01:27:53.280 --> 01:27:57.420 actions had on the system under observational control. 01:27:58.260 --> 01:28:03.240 And so you can evaluate whether that was the desired impact, or 01:28:03.240 --> 01:28:06.900 whether something different happened, and in this way you can evaluate 01:28:06.900 --> 01:28:10.920 the adequacy of a certain action that you selected in a certain 01:28:10.920 --> 01:28:15.120 situation, and it may be that you observe something which is different 01:28:15.120 --> 01:28:19.980 from what you knew before, and then you would modify, using this loop, 01:28:20.100 --> 01:28:22.160 you would modify the mapping adequately. 01:28:23.040 --> 01:28:27.400 Changing the information on the adequacy of an action in a certain 01:28:27.400 --> 01:28:27.980 situation. 01:28:29.420 --> 01:28:33.660 And then it may be that you have a situation, for example, we designed 01:28:33.660 --> 01:28:37.000 a traffic light controller with this architecture. 01:28:38.040 --> 01:28:41.980 Now, if you say I would like to find out whether it is reasonable to 01:28:41.980 --> 01:28:46.720 use a certain setting for the traffic light, and how the environment 01:28:46.720 --> 01:28:50.440 responds to that, I don't want to see car crashes on the intersection. 01:28:50.440 --> 01:28:55.440 So, there are safety requirements which might make it necessary to 01:28:55.440 --> 01:28:58.880 evaluate potential actions with respect to this model. 01:28:58.880 --> 01:29:04.600 So, you just run a simulation, you model certain actions, and then you 01:29:04.600 --> 01:29:10.060 find out what is, with respect to the model, the best possible 01:29:10.060 --> 01:29:15.940 activity or action for a certain situation, and then you can extend 01:29:15.940 --> 01:29:21.040 your mapping appropriately after you have run that simulation and some 01:29:21.040 --> 01:29:22.780 kind of optimization algorithm. 01:29:22.780 --> 01:29:29.480 You can, again, modify the mapping with more intelligent information 01:29:29.480 --> 01:29:33.140 on what kind of actions it should take in a certain situation where 01:29:33.140 --> 01:29:35.780 maybe you didn't have adequate responses before. 01:29:36.520 --> 01:29:38.420 So, this is the controller. 01:29:39.240 --> 01:29:43.300 This is doing mapping, evaluation, and adaptation. 01:29:44.120 --> 01:29:48.780 And on top here, we have the interface to the external user, which is 01:29:48.780 --> 01:29:52.020 providing directives for what the system should do. 01:29:52.760 --> 01:29:57.940 So, for example, for the traffic light scenario, possible directives 01:29:57.940 --> 01:30:03.120 are to reduce or to minimize the waiting time of the cars or the 01:30:03.120 --> 01:30:08.560 travel time of the cars through an intersection, or minimize the 01:30:08.560 --> 01:30:16.360 number of stops in the system, or minimize power or energy usage, 01:30:17.320 --> 01:30:19.400 carbon dioxide emissions, and things like that. 01:30:21.260 --> 01:30:24.320 So, this... I showed you that already. 01:30:25.100 --> 01:30:26.380 I should have taken that out. 01:30:27.640 --> 01:30:28.940 This is not... Okay. 01:30:28.940 --> 01:30:35.500 So, we can also have a different view on that architecture where we 01:30:35.500 --> 01:30:39.600 say we have some lower level, which is some kind of system which is 01:30:39.600 --> 01:30:45.380 running, having some specific type of control, and then we have the 01:30:45.380 --> 01:30:50.280 observer and controller on this first layer where we have a selection 01:30:50.280 --> 01:30:55.200 of different rules, that's the mapping, and some kind of adaptation of 01:30:55.200 --> 01:30:56.460 those rules. 01:30:57.660 --> 01:31:03.880 And if we don't have sufficient action knowledge available, that rule 01:31:03.880 --> 01:31:09.520 set, we go to the layer above, and then there we have the possibility 01:31:09.520 --> 01:31:12.650 to generate new rules using an evolutionary algorithm. 01:31:13.910 --> 01:31:19.810 And so this actually was designed for traffic light control, and we 01:31:19.810 --> 01:31:25.010 were capable to start it with almost no... or just with a simple 01:31:25.010 --> 01:31:30.630 traffic light controller, and very fast, it would come up with a very 01:31:30.630 --> 01:31:34.150 intelligent traffic light controller which would outperform a traffic 01:31:34.150 --> 01:31:34.550 engineer. 01:31:35.250 --> 01:31:36.770 So that was very successful. 01:31:37.810 --> 01:31:40.530 And here we have two kinds of loops. 01:31:41.070 --> 01:31:46.790 We have this online loop where we modify the... or where we adapt the 01:31:46.790 --> 01:31:51.870 systems based on what we observe in the system, and we have the 01:31:51.870 --> 01:31:58.450 offline loop where we modify the rule set based on what we optimize 01:31:58.450 --> 01:32:00.730 with respect to the simulator. 01:32:02.070 --> 01:32:09.870 And so then we can use those architectures in different ways. 01:32:09.970 --> 01:32:13.990 We can use it in a centralized way, as you see here. 01:32:14.970 --> 01:32:18.290 That is just one observer and controller. 01:32:18.290 --> 01:32:22.990 We could have a distributed architecture, which is shown here, where 01:32:22.990 --> 01:32:26.710 you have many such devices which could communicate in some way and 01:32:26.710 --> 01:32:28.510 optimize their behavior. 01:32:29.230 --> 01:32:32.790 And you could have a multilabel system, and that's actually what we 01:32:32.790 --> 01:32:37.110 assume to have in realistic organic computing systems. 01:32:37.110 --> 01:32:41.370 And what you also saw in the energy management system, the organic 01:32:41.370 --> 01:32:47.050 smart home, there we had one central observer inside the house, and 01:32:47.050 --> 01:32:49.510 all those observers for the individual devices. 01:32:50.770 --> 01:32:54.970 You can come up with many different other scenarios based on that 01:32:54.970 --> 01:33:00.350 technology, but this is essentially what... or just giving you some 01:33:00.350 --> 01:33:02.470 idea about what can be done. 01:33:03.390 --> 01:33:06.790 So, one last slide about organic computing. 01:33:07.050 --> 01:33:10.670 It has been applied to a range of different systems, different 01:33:10.670 --> 01:33:15.970 scenarios, and was a very successful priority research program of the 01:33:15.970 --> 01:33:21.630 German Research Foundation, and next time we will come back to the 01:33:21.630 --> 01:33:24.310 architecture for the energy management system. 01:33:24.810 --> 01:33:25.810 Thank you for your attention. 01:33:26.190 --> 01:33:27.010 That's it for today.