WEBVTT 00:00.890 --> 00:06.650 Okay, let's continue at the point where we stopped just... oops... the 00:06.650 --> 00:08.110 point where we just stopped before. 00:08.930 --> 00:15.410 So, just on the break, there was an interesting discussion on problems 00:15.410 --> 00:21.450 in this program of study, which we have to think about, and I really 00:21.450 --> 00:24.270 ask you to provide us the feedback. 00:24.650 --> 00:28.650 I assume that you have done that already to some people, but send your 00:29.310 --> 00:34.510 feedback on problems in the course just to the program coordinators, 00:35.190 --> 00:39.850 and then this will hopefully lead to some improvements. 00:40.630 --> 00:46.210 Okay, so I hope that the material that I am presenting to you is of 00:46.210 --> 00:47.130 some value for you. 00:47.470 --> 00:53.130 So, I'm really also looking for feedback from you. 00:53.930 --> 00:58.650 By the way, did we already... no, we did not ask for evaluation of the 00:58.650 --> 00:58.970 course. 00:59.150 --> 01:00.010 We should do that. 01:02.850 --> 01:09.450 We should take, next time maybe, we should take the evaluation forms 01:09.450 --> 01:12.570 and let them give some feedback. 01:14.630 --> 01:15.910 Can you organize that? 01:16.790 --> 01:17.690 Okay, thank you. 01:18.810 --> 01:21.010 So, let's come back to the course. 01:21.350 --> 01:23.370 Back to optimization of load profiles. 01:23.470 --> 01:28.650 I showed you the basic idea that we would like to take all these 01:28.650 --> 01:32.310 different schedules here, these different schedule or these different 01:32.310 --> 01:36.370 possibilities, and then combine them adequately into that area. 01:36.990 --> 01:41.130 And as we have seen, maybe that it doesn't fit. 01:41.790 --> 01:43.190 We have seen how we could do that. 01:43.450 --> 01:47.950 Use just some, like, make a randomized selection using evolutionary 01:47.950 --> 01:54.910 algorithms and optimize these, some kind of bin packing problems. 01:56.030 --> 01:58.590 And then I said we also use local search. 01:58.690 --> 02:03.510 Local search was not part of what I had in the ingredients of... oops, 02:03.810 --> 02:04.510 what did I do here? 02:05.410 --> 02:06.450 I didn't want to do that. 02:08.090 --> 02:12.230 I just wanted to select. 02:12.230 --> 02:19.070 Okay, so local search is something which you can do at any point in 02:19.070 --> 02:20.070 the evolutionary process. 02:20.810 --> 02:26.010 If you have, like, we had these different components here in our 02:26.010 --> 02:31.590 evolutionary algorithm, as you remember, and you could just put in 02:31.590 --> 02:33.170 here local search at some point. 02:33.510 --> 02:35.010 In every generation you could do that. 02:35.510 --> 02:42.070 That would mean you improve the current individuals in some way, look 02:42.070 --> 02:45.050 for better individuals in the neighborhood. 02:47.750 --> 02:52.170 In some way this is contradicting the idea of evolutionary algorithms, 02:52.330 --> 02:56.690 because in evolutionary algorithms we also say if you have a certain 02:56.690 --> 03:01.370 landscape, for example, a landscape like this, if you would like to 03:01.370 --> 03:06.890 minimize something, if you have something like this and there, and 03:06.890 --> 03:11.510 then you go actually come up something there, you would like to get to 03:11.510 --> 03:16.370 that position, but initially you are somewhere over here, and then 03:16.370 --> 03:22.150 maybe by the evolutionary process you actually get over that point 03:22.150 --> 03:22.330 here. 03:22.350 --> 03:23.830 You would like to move into that area. 03:24.490 --> 03:29.970 But if you're always looking for the local optimum, you would just 03:29.970 --> 03:32.510 never get across that point. 03:33.850 --> 03:38.150 In randomized search, in evolutionary algorithms, you might also end 03:38.150 --> 03:42.930 up by mutation or things like that at bad solutions, and then maybe 03:42.930 --> 03:46.570 have a chance to get into this area where you have the global optimum. 03:47.610 --> 03:56.790 So, in some way we rely on the potential to actually move into regions 03:56.790 --> 04:01.450 or to move with some individuals in regions where they are not 04:01.450 --> 04:07.410 performing very well, but have a chance to reach the really good areas 04:07.410 --> 04:08.090 from there. 04:09.150 --> 04:13.670 Nevertheless, it has turned out to be beneficial to actually perform 04:13.670 --> 04:14.310 local search. 04:14.410 --> 04:20.150 Something like this actually is called a mimetic algorithm. 04:21.530 --> 04:22.770 Mimetic algorithms. 04:22.890 --> 04:27.810 That's just a term characterizing those types of evolutionary 04:27.810 --> 04:29.690 algorithms where you have local search in between. 04:30.090 --> 04:35.410 Now, local search can mean if, for example, here we have an individual 04:35.410 --> 04:42.070 which might fit, but it's coming a little bit too late, then we could 04:42.070 --> 04:46.790 try to just shift it in time, and now it fits. 04:47.810 --> 04:48.770 Maybe possible. 04:49.490 --> 04:53.730 Or we see in the solution that we have in the evolutionary 04:53.730 --> 05:01.010 optimization, we have something which is not running long enough, and 05:01.010 --> 05:02.410 we just increase... 05:02.410 --> 05:03.610 Oops, that was the wrong time. 05:04.210 --> 05:06.110 The second year, later start time. 05:06.210 --> 05:09.290 So, that was earlier start time. 05:09.410 --> 05:13.370 The second is move something to a later start time. 05:14.910 --> 05:19.790 If we have the option, if it's possible to move something, to move the 05:19.790 --> 05:22.790 start time back or forth slightly, then we could do that. 05:23.510 --> 05:29.790 Maybe we can just let this device run a little bit longer, so increase 05:29.790 --> 05:30.270 the runtime. 05:30.790 --> 05:32.870 For some of them, it might be possible. 05:33.230 --> 05:38.290 Maybe that the evolutionary optimization did not really provide us 05:38.290 --> 05:43.830 with an element that fits exactly, but now we use local search like 05:43.830 --> 05:44.150 that. 05:44.410 --> 05:50.350 Or we decrease the runtime, so we would just stop it earlier. 05:50.990 --> 05:57.870 Or, you remember, there have been some potential loads which would 05:57.870 --> 06:00.290 vary in the power that they offer. 06:01.170 --> 06:06.010 So, we could increase the consumption, power consumption, or we could 06:06.010 --> 06:07.350 decrease power consumption. 06:08.890 --> 06:14.430 Based on the current contents of our solution, we could try to improve 06:14.430 --> 06:14.910 it locally. 06:15.610 --> 06:20.970 And then we might run up, or we might actually end up in a situation 06:20.970 --> 06:26.210 which is much better and fits the desired area completely. 06:26.930 --> 06:31.810 So, this is the idea to have just local search like that, local 06:31.810 --> 06:34.610 improvements, and in this way have a better solution. 06:36.410 --> 06:39.470 Now, this is what we actually would like to do. 06:40.030 --> 06:49.150 So, the idea is that we have a certain deviation and we would like to 06:49.150 --> 06:55.750 cope with that deviation at the next time instance by moving the 06:55.750 --> 06:57.270 demand to later times. 06:58.070 --> 07:04.390 Yeah, we move it to later times and then this should be evened out 07:04.390 --> 07:12.370 appropriately by other devices which would agree on actually using or 07:12.370 --> 07:17.770 producing more power and then would actually have a good schedule 07:17.770 --> 07:18.130 again. 07:19.730 --> 07:22.570 But you could do this by individual negotiations. 07:23.670 --> 07:30.850 You could also do it by, as I, oops, if you look at that situation 07:30.850 --> 07:31.270 again. 07:32.630 --> 07:37.350 We know that we have this extra demand here. 07:37.630 --> 07:41.390 We would like to shift that to a later time and this has to be a 07:41.390 --> 07:43.370 combined effect of several others. 07:43.950 --> 07:49.750 Then one individual, this one, or maybe the balancing group manager 07:49.750 --> 07:59.170 notices that and asks for the current potential flexibility of all the 07:59.170 --> 08:05.090 devices, runs the evolutionary algorithm and sends back the result, 08:05.750 --> 08:12.050 the selection of starting times for all the devices which offered some 08:12.050 --> 08:12.590 flexibility. 08:13.430 --> 08:17.750 And then the result of that optimization might be the schedule that is 08:17.750 --> 08:19.270 indicated here. 08:20.990 --> 08:25.850 So this is some kind of negotiation between the individuals but it's 08:25.850 --> 08:35.210 not shouting or asking everybody, can you move your load into some 08:35.210 --> 08:35.930 other place? 08:36.090 --> 08:38.650 They all agree to do that and you have an avalanche effect. 08:39.090 --> 08:45.370 But it is asking for the potential flexibility, collecting all that, 08:46.010 --> 08:53.170 optimizing based on the combined flexibility and then sending back the 08:53.170 --> 08:56.690 results and that gives the next schedule. 08:57.710 --> 09:02.430 So it is preventing something which might have avalanche effects and 09:02.430 --> 09:08.090 things like that because you would actually coordinate the shifting of 09:08.090 --> 09:10.450 demand and supply. 09:13.090 --> 09:18.530 So this is one possibility to do that but it means that it takes some 09:18.530 --> 09:20.850 time before you can actually come up with a solution. 09:20.970 --> 09:21.590 You have to optimize. 09:23.290 --> 09:26.630 If this is a large pool, it might really take some time before you 09:26.630 --> 09:27.690 have the optimum solution. 09:29.550 --> 09:34.430 This actually has been simulated on different settings. 09:35.010 --> 09:40.410 Actually in a setting where you have a virtual district of components, 09:40.810 --> 09:47.170 so about 2000 refrigerators and freezers, 84 combined heat and power 09:47.170 --> 09:55.410 plants, a simulation model running for one day, average consumption in 09:55.410 --> 10:01.790 that district 56 kilowatt hours per hour, so a load of 56 or power 56 10:01.790 --> 10:09.470 kilowatt over an hour, average production per hour 54 kilowatt hours. 10:09.610 --> 10:12.390 So this is balanced on the average. 10:13.610 --> 10:17.070 Then you certainly have spontaneous deviations and you have to cope 10:17.070 --> 10:19.050 with those deviations using that device. 10:20.090 --> 10:25.030 So you can have different kinds of imbalance. 10:28.750 --> 10:35.030 So here for example you have zero percent imbalance because here 10:35.030 --> 10:38.010 everything is eved out, demand and supply. 10:38.530 --> 10:42.730 Here you have 100 percent imbalance because you have the time where 10:42.730 --> 10:47.550 you have demand and no supply and here you have supply and no demand. 10:49.050 --> 10:50.630 Now that's very bad. 10:51.810 --> 10:56.470 And now we have to look at the percentage of imbalance and we will 10:56.470 --> 10:57.080 look at 11:02.590 --> 11:08.510 some scenario where we actually will end up having, without 11:08.510 --> 11:14.730 coordination, we have an imbalance of 11.21 percent. 11:15.190 --> 11:19.630 So some positions in the schedule we have an imbalance and with 11:19.630 --> 11:20.890 coordination this is reduced. 11:22.310 --> 11:26.310 And let's look at what kind of things in the simulation have been 11:26.310 --> 11:26.610 done. 11:27.650 --> 11:35.790 So assume we have just some balanced schedule and now we impose an 11:35.790 --> 11:37.470 external imbalance. 11:37.470 --> 11:41.390 We impose an increased demand at certain point of time. 11:41.930 --> 11:44.990 We impose an increased supply at this point in time. 11:45.750 --> 11:48.350 And we impose an increased demand again here. 11:49.250 --> 11:52.590 And the resulting schedule from that will look like this. 11:53.330 --> 12:00.350 The increased demand will lead to exactly that accumulated profile. 12:01.330 --> 12:05.430 And now we have to try to balance that again by shifting demand and 12:05.430 --> 12:05.710 supply. 12:06.210 --> 12:10.170 So we would have to shift something from here into that and so on. 12:10.870 --> 12:16.390 This shifting should be done by negotiating between those components, 12:16.630 --> 12:21.010 optimizing based on the available flexibility. 12:22.150 --> 12:28.990 This has been run in simulations and so it has been run on different 12:28.990 --> 12:36.590 scenarios where we have an imbalance of different sizes of 10 12:36.590 --> 12:39.870 kilowatt, 20, 30 or up to 60 kilowatt imbalance. 12:41.110 --> 12:48.250 And the duration for this imbalance occurs between 30 minutes and 90 12:48.250 --> 12:48.710 minutes. 12:49.550 --> 12:55.250 Then we see the resulting imbalance in the system if we have no 12:55.250 --> 12:57.770 coordination and if we have coordination. 12:57.950 --> 13:05.010 You see that certainly the imbalance is increased with the external 13:05.010 --> 13:06.490 imbalance that we impose. 13:07.450 --> 13:10.550 And if we have coordination we can reduce that. 13:10.850 --> 13:17.390 But still there's quite some remaining imbalance which is not really 13:17.390 --> 13:18.530 satisfying. 13:19.090 --> 13:26.110 So it means that we need essentially 25.93 percent in this example 13:26.750 --> 13:30.670 extra balancing power to get a balanced scheme. 13:32.170 --> 13:36.190 So this is showing that we can achieve something but it's not 13:36.190 --> 13:41.790 completely satisfying because we would like to reduce the need for 13:41.790 --> 13:44.810 extra balancing power even further. 13:47.190 --> 13:51.210 So what are the advantages of the device pool? 13:51.210 --> 13:53.390 We can reduce the imbalances. 13:54.130 --> 13:58.350 We don't need really administration of that because this is done by 13:58.350 --> 14:00.530 the mechanism in the pool that I described. 14:00.650 --> 14:06.410 We have the optimization process which can be run on, for example, the 14:06.410 --> 14:08.310 routers in the houses. 14:09.610 --> 14:11.930 It's not that difficult actually to run that. 14:12.550 --> 14:21.690 It's easily evaluated and we can use all the degrees of freedom in the 14:21.690 --> 14:26.650 system without having actually any bad side effects. 14:27.390 --> 14:34.070 The disadvantages are that essentially the imbalances have to be... we 14:34.070 --> 14:38.070 need this optimization and we have to provide the computing 14:38.070 --> 14:40.070 infrastructure for doing that. 14:40.190 --> 14:47.230 Maybe we can do everything on the routers but this is something which 14:47.230 --> 14:48.290 has to be looked at. 14:48.850 --> 14:51.210 The neighborhoods have to be set appropriately. 14:52.770 --> 14:56.390 So what are actually the neighborhoods of certain devices? 14:56.650 --> 15:01.070 What are the components which are potentially contributing to 15:01.070 --> 15:03.690 balancing demand and supply? 15:04.930 --> 15:11.550 And also how do we know if we see an imbalance how long that will be 15:11.550 --> 15:11.890 there? 15:12.310 --> 15:16.510 If it's a power plant which is going out of service, we might know 15:16.510 --> 15:19.610 this will not be operating for the next two hours. 15:20.150 --> 15:26.430 But if it's, let's say, just a car which is driving away or somebody 15:26.430 --> 15:32.170 just pulled the plug, maybe it will put in the plug again very soon, 15:32.570 --> 15:34.370 but you don't know that exactly. 15:34.770 --> 15:39.870 So this is something which is not easily really computed. 15:40.830 --> 15:44.610 And the overall consumption and production here is assumed to be 15:44.610 --> 15:47.570 fixed, so it should be more flexible. 15:49.290 --> 16:00.110 And now we go into an idea which actually can be more dynamical, more 16:00.110 --> 16:00.530 adaptive. 16:01.150 --> 16:06.150 We create something which is some kind of an elite group. 16:07.070 --> 16:10.830 Now the elite group can perform much better than everybody else. 16:12.090 --> 16:17.970 And the idea is that an elite group can respond immediately to 16:17.970 --> 16:20.830 spontaneous imbalances. 16:22.030 --> 16:25.850 So let's look at the idea. 16:26.190 --> 16:27.710 This is the idea of an elite group. 16:29.270 --> 16:35.570 We have our pool, and here we have an elite group of some of those 16:35.570 --> 16:36.150 devices. 16:37.070 --> 16:42.870 And these are the devices that have the maximum flexibility. 16:44.090 --> 16:48.970 I can always determine those devices in a pool of elements which are 16:48.970 --> 16:52.910 the most flexible, which are just running, for example, and can be 16:52.910 --> 17:00.390 stopped, or which can be switched on at any point in time, which is 17:00.390 --> 17:00.790 reasonable. 17:01.650 --> 17:07.330 So those are the devices where I don't have to ask what are your 17:07.330 --> 17:11.690 potential flexibilities and then optimize, but where I can say, okay, 17:11.730 --> 17:16.810 if there is an imbalance, if the consumption should go down, that's 17:16.810 --> 17:18.730 the one which will do that immediately. 17:19.870 --> 17:27.470 So we have an always available listing of components which can be 17:27.470 --> 17:34.490 triggered immediately whenever there is an imbalance and modify the 17:34.490 --> 17:41.290 demand, modify the load, either reduce consumption or increase 17:41.290 --> 17:41.890 generation. 17:45.390 --> 17:49.750 Now it may be that this is not sufficient, but maybe that, so for 17:49.750 --> 17:56.430 example, this one is doing that, and then, oops, here, this is moved 17:56.430 --> 17:58.130 to some later point, so this is reduced. 17:59.330 --> 18:07.610 Now this should lead to some, what else is here, oops, yeah, so the 18:07.610 --> 18:11.350 next imbalances have to be evened out. 18:12.250 --> 18:20.390 We can, here, they are actually reducing their power generation, and 18:20.390 --> 18:26.890 then maybe this device can even reduce its power consumption even 18:26.890 --> 18:27.250 more. 18:27.450 --> 18:33.990 That one can do a little bit, and then some other devices can modify 18:33.990 --> 18:43.850 their power demand or power generation, and then some imbalance will 18:43.850 --> 18:44.350 remain. 18:45.350 --> 18:53.350 But since the next imbalances will be treated first, those remaining 18:53.350 --> 18:55.630 imbalances will occur at a later time. 18:56.730 --> 19:02.670 And these later imbalances will be sent to the pool which can 19:02.670 --> 19:05.430 negotiate on how to deal with them. 19:06.470 --> 19:12.770 And in this way, we have the elite group which performs some kind of 19:12.770 --> 19:20.230 primary control or primary balancing power, and here we have our 19:20.230 --> 19:28.370 secondary control, which is taking care afterwards, taking care of the 19:28.370 --> 19:29.470 imbalances afterwards. 19:31.290 --> 19:34.630 Now, how do you select such an elite group? 19:35.570 --> 19:40.910 This can always be done in a very simple way, because you know, you 19:40.910 --> 19:44.970 can always determine those elements which are the most flexible. 19:45.890 --> 19:54.510 And if a device has actually provided its flexibility for evening out 19:54.510 --> 19:59.950 some imbalances, then it will lose its flexibility and will leave the 19:59.950 --> 20:00.430 elite group. 20:01.180 --> 20:06.150 And another one, in the meantime, might have become more flexible and 20:06.150 --> 20:07.370 might now enter the group. 20:08.530 --> 20:14.730 And so this elite group is actually formed in a self-organized way. 20:14.930 --> 20:18.390 Always the best performing elements are in that elite group. 20:19.150 --> 20:26.630 They can immediately provide response to these deviations in power 20:26.630 --> 20:32.150 demand or supply, and only the remaining imbalance is sent to the pool 20:32.150 --> 20:38.210 in some kind of secondary control, a secondary balancing power. 20:39.330 --> 20:44.250 And in that way, the pool needs more time to decide on how to deal 20:44.250 --> 20:45.390 with those imbalances. 20:46.350 --> 20:48.010 The elite group can do it immediately. 20:48.770 --> 20:54.090 So we have some kind of primary and secondary balancing devices. 20:55.730 --> 21:00.450 Okay, so this is here again showing a little bit more of those things. 21:01.490 --> 21:08.810 The results are that now the balancing group manager only sends that 21:08.810 --> 21:16.330 information to this system, which is the hierarchical system of 21:16.330 --> 21:18.510 primary elite group and secondary pool. 21:19.750 --> 21:23.970 And then the system will completely autonomously respond to that. 21:25.550 --> 21:34.190 Certainly, this group here should also provide some information on the 21:34.190 --> 21:39.170 maximum flexibility that it can offer currently as a primary response 21:39.170 --> 21:41.110 and maybe secondary response. 21:42.910 --> 21:48.010 And in this way, the balancing group manager always has the best 21:48.010 --> 21:49.190 information about that. 21:50.610 --> 21:55.230 And this optimization within the pool can be done by rather small 21:55.230 --> 21:57.610 computers, for example, the internet routers. 21:58.370 --> 22:01.430 The communication overhead is not that large. 22:01.690 --> 22:04.790 It certainly depends on the size of the neighborhood you have. 22:04.910 --> 22:09.730 Assume you have something like 80 to 100 devices in your neighborhood. 22:12.110 --> 22:17.390 There are many ideas or many things you could look at here. 22:18.210 --> 22:20.270 How would you actually set up the neighborhood? 22:20.490 --> 22:24.230 How would you decide which components should be grouped into one pool? 22:25.590 --> 22:27.330 They should complement each other. 22:27.450 --> 22:31.550 They should have properties which allow you to provide a more stable 22:31.550 --> 22:34.110 response to imbalances and things like that. 22:34.490 --> 22:38.010 So there are still quite a few problems which have to be addressed 22:38.010 --> 22:38.370 here. 22:38.750 --> 22:44.130 But the major idea is to use that elite pool and that secondary pool 22:44.130 --> 22:45.750 or the elite group and the pool. 22:46.470 --> 22:51.610 And then there is quite some potential to have a self-organizing 22:51.610 --> 22:55.990 system which will deal with local imbalances, which means that the 22:55.990 --> 23:01.810 need for extra external balancing power is or can be reduced. 23:03.590 --> 23:08.930 And this actually is protected by a pattern. 23:09.890 --> 23:11.570 This idea of doing that. 23:11.950 --> 23:16.650 But there are still quite a few open points if you put that into 23:16.650 --> 23:17.150 reality. 23:18.230 --> 23:23.430 So if you're interested in writing a thesis, you're welcome to do 23:23.430 --> 23:24.590 something on these topics. 23:26.430 --> 23:29.330 This is showing it in a slightly different way. 23:30.270 --> 23:32.550 Controlled self-organizing energy management. 23:33.050 --> 23:35.850 We have our elite group. 23:36.570 --> 23:40.890 In every elite group we have some kind of observer and controller. 23:42.450 --> 23:45.650 So a local, some kind of feedback control system. 23:46.950 --> 23:50.430 And they communicate with the pool of devices. 23:51.730 --> 23:55.690 That would be one elite group for a pool of those devices which all 23:55.690 --> 23:56.990 know what they can do. 23:57.680 --> 24:02.450 They can respond on what they get, whatever information they get from 24:02.450 --> 24:02.970 the others. 24:03.730 --> 24:09.150 Maybe that you have several elite groups communicating with certain 24:09.150 --> 24:09.910 components. 24:12.190 --> 24:18.990 Then those like external information on the amount of energy that 24:18.990 --> 24:23.430 currently should be made available. 24:23.550 --> 24:30.710 Extra load maybe for offering it on the market or for other purposes. 24:31.250 --> 24:34.490 Can be split over several elite groups. 24:35.470 --> 24:41.970 The balancing group manager could send or split the demand for load 24:41.970 --> 24:45.750 over several elite groups knowing what they can provide. 24:45.930 --> 24:51.490 Then he just splits the demand for energy over to these different 24:51.490 --> 24:58.650 groups and how that is implemented is something which is not the 24:58.650 --> 25:02.490 concern of the balancing group manager is organized in a self 25:02.490 --> 25:04.210 -organized way internally. 25:04.930 --> 25:10.890 So this is the idea behind that and that decentralized system 25:10.890 --> 25:12.870 operating autonomously. 25:13.950 --> 25:18.410 And the effects of that of those tools within the lead group are shown 25:18.410 --> 25:19.670 here in these diagrams. 25:20.470 --> 25:23.970 So here we have several different diagrams. 25:24.790 --> 25:31.410 Here we have a certain external imbalance which is shown in a very 25:31.410 --> 25:32.170 regular way. 25:32.250 --> 25:37.570 The green curve here is showing regular imbalance positive and 25:37.570 --> 25:38.550 negative load. 25:40.070 --> 25:43.550 And you have the generation. 25:43.910 --> 25:45.530 You have the consumption. 25:46.630 --> 25:52.910 And here you actually see the deviation that is resulting. 25:53.090 --> 25:54.230 You see there's no deviation. 25:55.450 --> 26:01.090 So the system on such a regular external imbalance in those 26:01.090 --> 26:06.170 simulations the systems actually could respond perfectly and you had 26:06.170 --> 26:09.750 almost zero imbalance in the system. 26:10.670 --> 26:15.470 Now the external imbalance can be very different. 26:16.670 --> 26:19.010 Then you have to respond to that. 26:19.870 --> 26:28.410 And even here the result was that the the demand and supply of energy 26:28.410 --> 26:35.890 adjusted to that even to such a more random pattern of external 26:35.890 --> 26:39.250 imbalances, external loads. 26:40.770 --> 26:52.670 And actually this is showing the same for many more households. 26:52.930 --> 26:59.010 So here we actually have 20 elite groups on 20,000 households whereas 26:59.010 --> 27:04.610 this corresponded to 1,000 households and just having one elite group. 27:06.170 --> 27:12.250 And so here also the deviation the resulting deviation is almost zero 27:12.250 --> 27:15.210 just by using those elite groups. 27:16.050 --> 27:19.110 And certainly it depends on the kind of profiles you assume for the 27:19.110 --> 27:19.930 different devices. 27:20.330 --> 27:24.030 But these devices were freezers, fridges, combined heat and power 27:24.030 --> 27:29.110 plants using typical profiles and flexibilities of those elements. 27:29.790 --> 27:35.090 It shows that principle using this approach you can actually reduce 27:35.090 --> 27:39.630 the demand for balancing power almost to zero, just locally. 27:40.490 --> 27:45.330 Whereas before we had seen that it would lead to quite some extra 27:45.330 --> 27:49.130 demand for balancing power. 27:50.390 --> 27:53.250 Okay, so that is the effect of this approach. 27:53.370 --> 27:58.450 A very decentralized approach working in a self-organized way. 27:59.250 --> 28:03.070 And so this is true decentralized energy management. 28:03.070 --> 28:08.930 So the summary of that chapter is that self-organizing balancing power 28:08.930 --> 28:12.890 is a promising approach for reducing the need for balancing power. 28:15.910 --> 28:21.250 In spite of the positive results that I show here, these results are 28:21.250 --> 28:22.170 based on simulations. 28:23.290 --> 28:28.790 And in practical applications we still have to investigate that in 28:28.790 --> 28:33.970 particular this capability to respond in real time. 28:35.810 --> 28:40.030 Yeah, does it actually work that we cause that we respond with these 28:40.030 --> 28:44.330 devices even from the elite group in real time within the time 28:44.330 --> 28:48.330 constraints that we have for immediate response. 28:49.710 --> 28:56.090 And another open issue is the derivation of guarantees, service levels 28:56.090 --> 28:58.230 for the duration of local balancing power. 28:58.970 --> 29:05.070 So these devices are randomly selected devices in some way. 29:05.590 --> 29:09.750 So how can we come up with service levels if you are offering such a 29:09.750 --> 29:16.410 pool of devices and you would like to perform or to transform that 29:16.410 --> 29:17.610 into some business model? 29:18.910 --> 29:23.570 You invest something on running that intelligent software there, how 29:23.570 --> 29:24.650 can you make money from that? 29:25.130 --> 29:27.090 You are offering a certain capability. 29:27.950 --> 29:31.030 You are offering a certain capability to the balancing group manager. 29:32.110 --> 29:34.130 Now you would like to be paid for that. 29:34.570 --> 29:38.110 But then you also have to give guarantees that you can achieve a 29:38.110 --> 29:40.670 certain level of balancing. 29:41.990 --> 29:49.330 Now how can you get reliable or dependable guarantees based on the 29:49.330 --> 29:50.730 properties of your devices? 29:51.050 --> 29:56.290 How do you optimize your pool such that you can best provide these 29:56.290 --> 29:57.330 guarantees? 29:58.030 --> 30:01.970 These are problems that have to be looked at so there's quite a bit of 30:01.970 --> 30:03.850 room for further improvement. 30:05.650 --> 30:08.550 That's what I wanted to present to you about the self-balancing 30:08.550 --> 30:09.850 approach. 30:11.030 --> 30:19.590 Okay, then we get to the next chapter, which should be chapter four, 30:21.250 --> 30:23.610 smart home and energy management. 30:26.010 --> 30:27.270 It's not completely 30:30.530 --> 30:34.010 done, but most certainly for today. 30:34.630 --> 30:37.650 Do you have questions with respect to the things that I presented? 30:49.460 --> 30:50.100 Yes. 30:52.640 --> 30:55.580 Yes, you're completely right. 30:55.720 --> 30:57.920 And we will come exactly to that question in this chapter. 30:58.880 --> 31:00.240 Yeah, how can we actually deal? 31:01.440 --> 31:08.160 These washing machines were also included in that pool, but how they 31:08.160 --> 31:13.200 can know about the flexibility that they have is not that trivial. 31:14.020 --> 31:21.620 And we have to find out how we can get that information from the owner 31:21.620 --> 31:22.700 of the washing machine. 31:23.640 --> 31:25.660 And there are simple ways of doing that. 31:26.280 --> 31:27.520 And we'll show you how we do that. 31:28.640 --> 31:29.660 Other questions? 31:31.880 --> 31:36.440 So please feel free to just put up your hands or just ask a question. 31:37.020 --> 31:43.300 I always have the tendency of just telling you many things and I 31:43.300 --> 31:44.420 always look at you. 31:44.560 --> 31:47.860 That's the advantage of this way of presenting things that I don't 31:47.860 --> 31:49.960 have to write on the blackboard, but I write on the screen. 31:50.060 --> 31:52.660 And so I can all the time look at you and see at your responses. 31:53.260 --> 31:55.700 But I'm interested in getting responses from you. 31:55.780 --> 31:58.060 So if you have questions, please feel free to ask. 31:59.560 --> 32:04.840 Okay, so let's look at smart home and energy management. 32:06.800 --> 32:14.300 So I start with a slide where I show you some, or there has been some 32:14.300 --> 32:19.700 funding program in Germany on this type of research that I will 32:19.700 --> 32:20.400 present to you. 32:20.420 --> 32:27.380 So the energy program was a program which had the objective of getting 32:27.380 --> 32:31.920 solutions, getting technologies where energy technology is combined 32:31.920 --> 32:35.900 with market mechanisms and information communication technology in 32:35.900 --> 32:38.400 order to improve the system. 32:39.640 --> 32:47.000 And we had another funding program on ICT for electric mobility, on 32:47.000 --> 32:51.960 model regions for electric mobility, and on research on electrical 32:51.960 --> 32:52.440 storage. 32:52.600 --> 32:54.420 So a lot of research is being done here. 32:55.240 --> 32:59.680 And I will present to you briefly what we did actually here, what kind 32:59.680 --> 33:06.140 of concepts have originated from two projects that we looked at. 33:06.600 --> 33:11.900 Project Meragio moving towards minimum emission regions, that was done 33:11.900 --> 33:13.220 within the energy program. 33:13.780 --> 33:18.440 The project Meragio Mobile, that was done in the funding program ICT 33:18.440 --> 33:24.460 for electric mobility, which is meanwhile extended with a program on 33:24.460 --> 33:28.260 over the project called Chrome, cross-border mobility with electric 33:28.260 --> 33:33.900 vehicles, and ITSOIS, which is called Intelligent Zero Emission Urban 33:33.900 --> 33:34.420 Systems. 33:34.840 --> 33:39.240 So they all have to do with electric mobility, with designing 33:39.240 --> 33:45.940 intelligent services for dealing with the need for power for energy in 33:45.940 --> 33:46.940 those electric vehicles. 33:48.500 --> 33:53.340 And the first one here, Meragio, is on smart home energy optimization. 33:54.700 --> 33:56.560 This was spread all over Germany. 33:57.640 --> 34:03.240 And so KIT is involved, has been involved in Meragio and Meragio 34:03.240 --> 34:03.540 Mobile. 34:04.120 --> 34:06.900 Oops, those are certainly connected. 34:08.020 --> 34:11.000 And so what did we do in Meragio? 34:11.740 --> 34:16.720 As I said, the idea was to use or to combine energy technology, energy 34:16.720 --> 34:19.040 markets, and information and communication technology. 34:19.760 --> 34:22.700 So what kinds of energy technology do we have? 34:22.760 --> 34:27.540 We have smart metering, we have all kinds of generation of power, we 34:27.540 --> 34:31.220 have demand side management, distribution grid management, there are 34:31.220 --> 34:36.480 energy markets, market mechanisms that we have in the European Energy 34:36.480 --> 34:42.620 Exchange, might be of interest to have regional markets or very 34:42.620 --> 34:43.840 decentralized markets. 34:45.100 --> 34:48.680 We have information and communication technologies, how can we use 34:48.680 --> 34:50.640 them to improve the energy systems. 34:51.580 --> 34:55.440 So our objectives here, these were the general objectives of the 34:55.440 --> 34:55.920 program. 34:56.480 --> 35:02.160 This is this device for showing or visualizing time varying prices. 35:03.080 --> 35:09.240 And we have the objectives of our projects to optimize power 35:09.240 --> 35:12.380 generation and usage from producers to end consumers. 35:13.040 --> 35:21.080 So how can we make sure that power is produced adequately based on the 35:21.080 --> 35:22.160 demand by consumers? 35:22.600 --> 35:28.060 And how can we combine new generator technology, demand side 35:28.060 --> 35:29.540 management, and ICT? 35:30.320 --> 35:31.720 What about price signals? 35:32.200 --> 35:34.120 What about combination of heat and power? 35:34.660 --> 35:40.500 And how can we actually make statements on the quality of a certain 35:40.500 --> 35:42.880 region with respect to energy management? 35:43.760 --> 35:49.180 That's called the Meridia Certificate, where we actually make a 35:49.180 --> 35:54.080 statement, well, this is an A plus or A plus plus region performing 35:54.080 --> 36:00.260 very well, and another region might be very bad on intelligent use of 36:00.260 --> 36:00.520 power. 36:01.720 --> 36:06.980 And in this we cooperated with the ENBW, the local energy company, the 36:06.980 --> 36:11.360 ABB, IBM, SAP, and a small consulting company, System Plan. 36:11.980 --> 36:18.440 They had to tell companies, in particular small and medium-sized 36:18.440 --> 36:21.500 companies, how they could take part in that. 36:22.080 --> 36:25.320 And we had this pilot region with 1,000 participants in the Black 36:25.320 --> 36:30.580 Forest, and five chairs at KIT from different disciplines, energy 36:30.580 --> 36:34.460 economics, informatics, telematics, management, and law. 36:35.540 --> 36:39.760 And this is actually what you should have in your program or study. 36:40.780 --> 36:42.480 Combination of those disciplines. 36:44.160 --> 36:48.520 Because that's what you need in order to build the new energy system. 36:50.060 --> 36:51.140 You expected those. 36:52.300 --> 36:56.800 But you get courses, you get classes in the area of energy economics. 36:59.080 --> 36:59.640 Hmm? 37:13.110 --> 37:14.810 Oh, really? 37:15.290 --> 37:17.430 But we have really good people who can offer that. 37:18.410 --> 37:18.550 Yeah? 37:19.790 --> 37:23.530 You have to give, like, we have to memorize that feedback. 37:23.990 --> 37:25.510 Informatics, just a little bit. 37:26.430 --> 37:31.150 I would have liked to offer you also a course on energy informatics, 37:31.970 --> 37:36.010 but my capacity is just limited. 37:36.870 --> 37:42.370 And so what I do, this is some kind of informatics background on, or a 37:42.370 --> 37:47.190 little bit at least, showing you how we use ICT in these areas. 37:47.710 --> 37:51.830 Telematics, you should know something about communication protocols 37:51.830 --> 37:54.810 and things like that, which is not part of this program. 37:55.610 --> 37:58.890 But we are cooperating with people from that area. 37:59.490 --> 38:03.030 They are not really included in the program. 38:03.990 --> 38:08.470 Management and law, the law people, they can tell us something about 38:08.470 --> 38:11.650 the regulatory requirements of the energy markets. 38:12.690 --> 38:15.630 And so you don't have courses on that either? 38:16.590 --> 38:17.190 No? 38:19.290 --> 38:23.650 So it's, there are deficiencies in that program. 38:23.790 --> 38:24.310 I see that. 38:25.150 --> 38:27.490 Okay, so what did we do in Meragio? 38:27.950 --> 38:31.270 In Meragio, we had four phases of the program. 38:31.650 --> 38:35.110 We had to set it up, finding out the model regions, and then 38:35.110 --> 38:37.610 installing initial technology. 38:38.010 --> 38:39.870 So we first looked at price elasticity. 38:40.270 --> 38:43.930 I showed you slides on that already, but we looked at the price 38:43.930 --> 38:50.030 elasticity, depending on, like, responses of people to these time 38:50.030 --> 38:50.730 -varying prices. 38:51.710 --> 38:56.730 We designed, there are certain concepts were developed, how we could 38:56.730 --> 38:59.410 control individual devices in households. 38:59.590 --> 39:03.350 Like, how we could control, actually, fridges or freezers, depending 39:03.350 --> 39:05.370 on the current situation in the energy market. 39:06.750 --> 39:08.490 840 people were doing that. 39:09.870 --> 39:13.570 But 840 more people than the initial 100 people. 39:14.130 --> 39:17.970 And then 40 of them were equipped with storage facilities, like 39:17.970 --> 39:24.570 batteries, which could also be used to be integrated into the system. 39:24.570 --> 39:28.230 And in particular, in the Friant region, they have a surplus of 39:28.230 --> 39:28.590 energy. 39:28.790 --> 39:32.430 They have a lot of photovoltaic systems, and wind power parks, and so 39:32.430 --> 39:32.630 on. 39:33.070 --> 39:36.710 And so, their storage facilities are really of interest. 39:37.350 --> 39:42.670 And then, they have all been given the possibility to play a little 39:42.670 --> 39:47.230 bit with market mechanisms, in a very simplified market, to see how 39:47.230 --> 39:49.770 they would actually operate on the market. 39:50.270 --> 39:58.850 So, this was about the four phases of Meragio, which ran from 2008 to 39:58.850 --> 39:59.850 2012. 40:01.430 --> 40:03.470 And I showed you this already. 40:04.430 --> 40:06.510 This changed the user responses. 40:06.690 --> 40:09.110 How can we, how can you actually measure user responses? 40:09.750 --> 40:11.550 You need a certain group. 40:11.770 --> 40:15.750 So, here we have two groups. 40:16.310 --> 40:20.930 We have the Meragio group, and we have a reference group. 40:21.630 --> 40:26.790 And it was measured like the typical demand that they have over a 40:26.790 --> 40:27.770 certain week. 40:29.010 --> 40:35.310 And then, after certain days, the demand curve, the demand profiles 40:35.310 --> 40:36.250 were looked at again. 40:37.690 --> 40:42.830 So, in this, after this time, those initial participants had been 40:42.830 --> 40:47.290 confronted with the visualization of the energy consumption and these 40:47.290 --> 40:48.250 varying prices. 40:49.010 --> 40:53.570 And the effect that was visible was what I showed you before already, 40:54.270 --> 40:59.950 the responses to green, yellow, and red price signals on different 40:59.950 --> 41:01.450 times of the day. 41:01.650 --> 41:07.270 So, the response on green signal to improve, to increase power 41:07.270 --> 41:09.130 consumption was obvious. 41:10.050 --> 41:14.670 And, interestingly, we actually noticed a decrease in power 41:14.670 --> 41:19.430 consumption just as an effect of visualizing the power consumption 41:19.430 --> 41:20.390 profile. 41:20.830 --> 41:23.710 Because then you get aware of what you are consuming. 41:24.370 --> 41:25.690 Say, oh, I can switch that off. 41:26.710 --> 41:29.930 Then you don't shift only, but you just switch it off, which certainly 41:29.930 --> 41:31.590 has an even better effect. 41:33.610 --> 41:39.370 And then, this was what I showed you already before, so I should have 41:39.370 --> 41:41.490 taken it out at the other place. 41:42.450 --> 41:46.330 So, there we, that's what we looked at in this project. 41:47.190 --> 41:51.710 I think I will, later on, I will show you a little more, a few more 41:51.710 --> 41:54.410 things that we looked at in the Meragio project. 41:54.410 --> 41:59.650 Like, we also looked at how we can use the information from the pilot 41:59.650 --> 42:05.550 region in order to optimize the demand, no, the distribution system 42:05.550 --> 42:08.110 bottlenecks. 42:08.270 --> 42:11.950 And, like, the distribution system operator got information from the 42:11.950 --> 42:17.690 model region and then would calculate or would predict the situation 42:17.690 --> 42:18.370 in the network. 42:18.850 --> 42:22.730 Whenever there are bottlenecks predicted, they would send that 42:22.730 --> 42:27.770 information with some kind of priority signal to modify the current 42:27.770 --> 42:31.590 load in the system and, in that way, prevent bottlenecks from 42:31.590 --> 42:31.970 occurring. 42:33.590 --> 42:38.690 Then we had this additional project, Meragio Mobile, where we looked 42:38.690 --> 42:43.070 at ICT for electric mobility, how we could actually integrate electric 42:43.070 --> 42:44.270 vehicles into the grid. 42:45.170 --> 42:48.710 So, certainly, there's one technological challenge, how we can 42:48.710 --> 42:55.150 actually design an electric car such that we can use electricity or 42:55.150 --> 42:58.250 transform electricity into a driving power. 42:59.270 --> 43:03.010 Then you need to actually provide that electricity. 43:03.610 --> 43:07.090 So, we have to talk about the integration of electric vehicles into 43:07.090 --> 43:07.470 the grid. 43:09.190 --> 43:12.570 And this should be looked at under real-life conditions. 43:12.710 --> 43:15.750 So, we had to, again, look at few tests. 43:16.790 --> 43:21.490 And we also were interested in integrating that into the Meragio pilot 43:21.490 --> 43:21.990 region. 43:22.770 --> 43:26.010 And in order to experiment with that, we built a center of competence 43:26.010 --> 43:27.910 at KIT, a demo and research lab. 43:27.950 --> 43:29.370 And we'll show you what we did there. 43:29.910 --> 43:33.570 And I promise you that we will go there sometime and have a look at 43:33.570 --> 43:34.750 what we actually have done. 43:34.930 --> 43:38.050 Although you'll see some slides where you see most of that. 43:38.570 --> 43:41.530 And in this program, we cooperated again with the ENBW. 43:41.710 --> 43:43.510 They were the coordinators of the project. 43:43.510 --> 43:48.690 We also had SAP in here, but then we had Daimler and Opel, Fraunhofer 43:48.690 --> 43:54.310 Research Institute, the municipal power supplier, Stadtwerke, Bosch, 43:54.370 --> 43:55.290 the automotive company. 43:56.510 --> 44:04.150 And this time we had 11 chairs from KIT, also people from electrical 44:04.150 --> 44:07.750 engineering, and a few more people from computer science. 44:09.230 --> 44:09.650 Okay. 44:10.310 --> 44:14.370 Now, what is so interesting about electric mobility with respect to 44:14.370 --> 44:15.130 the energy system? 44:16.310 --> 44:19.130 Like, this is showing the development plan that we have in Germany. 44:19.990 --> 44:20.990 This is already over. 44:21.110 --> 44:23.610 The first, oops, what is this? 44:25.630 --> 44:29.570 I didn't want to show you that. 44:30.650 --> 44:31.830 I wanted to stay here. 44:33.450 --> 44:37.290 So, this was the first phase. 44:38.530 --> 44:43.530 The market technology preparation was done in the projects from 2009 44:43.530 --> 44:44.430 to 2011. 44:44.890 --> 44:51.110 We had an economic incentive package, and we actually had many 44:51.110 --> 44:53.410 projects on those technologies. 44:53.790 --> 44:56.750 So, that's when we operated that, or when we started that project. 44:57.170 --> 44:58.470 Now, we are in this phase. 44:58.870 --> 45:01.890 Here, we have the I-Source project, and we have a leading-edge cluster 45:01.890 --> 45:07.010 electric mobility, and we have showcase projects, and so on. 45:07.570 --> 45:09.550 Later on, we will have the volume market. 45:09.830 --> 45:14.870 The goal for 2020, as you might know, is to have actually 1 million 45:14.870 --> 45:17.330 electric vehicles in 2020 in Germany. 45:19.370 --> 45:23.650 People are quite suspicious whether that actually will be achieved. 45:23.710 --> 45:24.310 You never know. 45:25.450 --> 45:29.930 Currently, it doesn't look like that we will actually achieve that, 45:30.250 --> 45:32.050 but you never know how industry is developing. 45:32.350 --> 45:37.690 Maybe they somehow, all of a sudden, produce and fabricate lots of 45:37.690 --> 45:42.490 electric vehicles, which are very cheap, which is very unlikely, but 45:42.490 --> 45:43.470 you never know. 45:44.530 --> 45:46.410 Unlikely things might happen sometimes. 45:47.390 --> 45:53.650 And the objective is to make Germany into the lead market and lead 45:53.650 --> 45:56.250 provider for electric mobility. 45:57.590 --> 45:58.550 Let's see what's happening. 45:58.650 --> 46:02.970 Currently, it doesn't look like it, but we are working on it. 46:03.550 --> 46:04.910 What do we need for that? 46:05.450 --> 46:07.670 We need battery technology, definitely. 46:07.930 --> 46:09.810 We need competence in that area. 46:10.730 --> 46:14.850 Actually, a few years before that, all research on batteries was 46:14.850 --> 46:15.550 stopped in Germany. 46:16.530 --> 46:20.790 There had been funding programs, but people said, we don't need 46:20.790 --> 46:21.110 batteries. 46:21.530 --> 46:23.550 So, they stopped all the research in that area. 46:23.950 --> 46:25.190 The federal funding was stopped. 46:26.310 --> 46:29.090 And then, all of a sudden, I noticed, oh, it's getting more 46:29.090 --> 46:34.010 interesting because we have all these fluctuating power suppliers from 46:34.010 --> 46:35.910 photovoltaic panels and wind and so on. 46:36.390 --> 46:38.150 Maybe batteries are a good idea. 46:39.030 --> 46:41.290 And so, now, everybody is doing research on that. 46:42.130 --> 46:47.290 Also, some years ago, all the chairs for electrical grids, electrical 46:47.290 --> 46:49.490 grid technology, were phased out. 46:50.270 --> 46:51.410 Grids are boring. 46:51.590 --> 46:55.290 We know how to transport energy and how to distribute energy. 46:56.290 --> 46:58.590 What kind of research questions do we have there? 46:59.570 --> 47:04.550 Now, everybody is installing new chairs on smart grids because, all of 47:04.550 --> 47:06.370 a sudden, the situation is completely changing. 47:07.570 --> 47:12.430 So, there really has to be done quite a bit of research on building 47:12.430 --> 47:13.550 the new infrastructure. 47:14.350 --> 47:18.290 So, we need an interoperable and large-scale charging infrastructure 47:18.290 --> 47:19.390 for electric vehicles. 47:20.270 --> 47:21.650 Many open problems there. 47:22.530 --> 47:24.530 I will tell you more about that a bit later. 47:26.190 --> 47:31.290 And we also need those battery electric vehicles or plug-in hybrid 47:31.290 --> 47:32.210 electric vehicles. 47:32.790 --> 47:36.970 So, maybe those one million electric vehicles will be mainly plug-in 47:36.970 --> 47:37.530 hybrids. 47:38.530 --> 47:39.730 More realistic. 47:40.670 --> 47:42.030 Or range extenders. 47:42.490 --> 47:43.610 Also realistic. 47:44.290 --> 47:48.590 But pure battery electric vehicles may be in cities and so on. 47:49.390 --> 47:53.850 So, these are the electric cars that will be developed in some way. 47:54.810 --> 47:59.830 The question is, what is their influence on the electric system? 48:00.310 --> 48:04.110 So, we have our electric vehicles. 48:04.310 --> 48:07.590 They are supposed to provide mobility services. 48:07.770 --> 48:11.610 So, let's look at what the typical mobility requirements are. 48:12.850 --> 48:18.430 And there are continuous panels on mobility in Germany, where people 48:18.430 --> 48:21.210 are asked about their mobility patterns. 48:21.890 --> 48:23.250 How they use their car. 48:24.430 --> 48:31.610 And in 2008, one sample year, it showed that the average daily car 48:31.610 --> 48:33.390 usage is less than one hour. 48:33.670 --> 48:35.690 Most of the time the cars are sitting around. 48:37.070 --> 48:42.490 You can make more interesting diagrams on where these times or the 48:42.490 --> 48:48.930 locations of the car are split between home and work and supermarket. 48:50.170 --> 48:51.630 And things like that. 48:52.330 --> 48:54.190 You can look at the length of the trips. 48:54.370 --> 49:00.830 And you will notice that, well, I think 50% of the trips are less than 49:01.990 --> 49:03.690 20 kilometers or so. 49:04.090 --> 49:08.330 94% of the trips are actually less than 50 kilometers. 49:09.430 --> 49:14.690 Now, one of the major arguments against electric mobility is, oh, they 49:14.690 --> 49:16.150 have such a limited range. 49:16.770 --> 49:18.650 Just 150 to 200 kilometers. 49:20.110 --> 49:21.530 You don't need that. 49:22.930 --> 49:26.570 50 kilometers is more than you need for 94% of your trips. 49:27.290 --> 49:32.470 Certainly, it's those 6% spontaneous trips that get you worried. 49:33.410 --> 49:39.770 Because maybe that I decide to go to Stuttgart and back, and then I 49:39.770 --> 49:43.550 don't know whether it actually will be sufficient, like whether the 49:43.550 --> 49:45.550 battery is sufficiently charged for that. 49:46.030 --> 49:47.770 Or I would like to go to Frankfurt. 49:48.290 --> 49:49.310 That doesn't work. 49:50.570 --> 49:52.270 So this is a problem. 49:53.010 --> 49:55.310 The range anxiety is a problem. 49:56.160 --> 50:00.370 So-called range anxiety. 50:00.790 --> 50:02.610 Buzzword for electric mobility. 50:04.170 --> 50:08.490 It's not really there, or it's not justified. 50:09.830 --> 50:13.130 For example, people have been asked in a few tests of electric 50:13.130 --> 50:16.710 mobility whether they would need public charging stations. 50:17.550 --> 50:20.750 And they all said, yes, we need public charging stations, because we 50:20.750 --> 50:25.610 will drive around, and we will need the capability to just recharge 50:25.610 --> 50:26.230 our battery. 50:27.170 --> 50:32.010 And after a year of using electric vehicles in larger cities, I think 50:32.010 --> 50:37.850 it was in London, nobody asked for public charging stations, because 50:37.850 --> 50:41.910 they noticed, oh, we can always get home on our daily trips, and we 50:41.910 --> 50:44.110 are recharging our batteries at home. 50:45.370 --> 50:50.070 But if you are used to driving a normal car with an internal 50:50.070 --> 50:54.870 combustion engine, then you go to get gas at a gas station. 50:54.910 --> 51:00.090 You don't get gas at home, but you get gas only at the gas station. 51:00.650 --> 51:05.570 And so being used to that pattern, you expect to get power at some 51:05.570 --> 51:06.690 public charging station. 51:07.210 --> 51:08.870 Then you notice, oh, I don't need that at all. 51:10.070 --> 51:14.830 And you will notice if you now have to decide as a politician, how can 51:14.830 --> 51:18.270 I get people into using electric vehicles? 51:19.110 --> 51:20.950 They will all ask for public infrastructure. 51:21.830 --> 51:24.530 But I know that it will not be needed in the end. 51:25.590 --> 51:29.030 So a simple idea would be just to put all kinds of charging stations 51:29.030 --> 51:31.290 there, and most of them are just fake. 51:32.610 --> 51:33.790 They will never be used. 51:34.970 --> 51:37.950 So here at Karlsruhe, we have 25 or 30 charging stations. 51:38.110 --> 51:41.530 They are all operational, but most of them are never used. 51:42.630 --> 51:44.350 Yeah, so it's an interesting point. 51:44.630 --> 51:46.230 But it's not the major point of this slide. 51:46.810 --> 51:52.970 So this is the major point here is that the cars are moving only for a 51:52.970 --> 51:55.310 fraction of the time during the day. 51:56.170 --> 51:59.030 And then we can look at electric vehicles. 51:59.410 --> 52:02.170 If we have an electric vehicle, it will have the battery. 52:03.210 --> 52:08.250 And let's assume that we have an average capacity in those batteries 52:08.250 --> 52:09.510 of 20 kilowatt hours. 52:10.470 --> 52:17.630 So the Meragio, the Opel Meriva Meragio, the electric vehicle that we 52:17.630 --> 52:23.310 have in our project, has something like 16 kilowatt hours. 52:23.430 --> 52:28.130 But in those that are on the market and that are assumed to get on the 52:28.130 --> 52:31.090 market, we even have something like 30 kilowatt hours. 52:31.310 --> 52:34.070 20 is a conservative assumption. 52:34.730 --> 52:36.010 20 kilowatt hours. 52:37.410 --> 52:40.990 But now assume that you have 1 million battery electric vehicles or 52:40.990 --> 52:44.630 electric vehicles having that capacity, which is our objective for 52:44.630 --> 52:45.230 2020. 52:46.530 --> 52:51.670 That's a combined storage capacity of 20 gigawatt hours, which is 52:51.670 --> 52:52.150 substantial. 52:53.650 --> 52:56.790 Yeah, now how can you use that, utilize that? 52:57.290 --> 52:59.170 You have to connect the batteries to the system. 52:59.230 --> 52:59.990 How do you do that? 53:00.650 --> 53:04.870 A simple way, you would just use single-phase charging. 53:05.210 --> 53:10.050 Just plug in your cable, use single-phase charging with 16 ampere, 53:10.470 --> 53:12.590 which means you have 3.7 kilowatt. 53:13.090 --> 53:17.890 If you would sum all that up for 1 million battery of 1 million cars, 53:17.970 --> 53:21.010 you would have 3.7 gigawatt potential power. 53:21.600 --> 53:28.190 Now this means 3.7 gigawatt additional load on the system. 53:28.690 --> 53:32.950 If they all want to be charged at the same time, it means you need 53:32.950 --> 53:38.810 that much extra power, which is about the power that three nuclear 53:38.810 --> 53:40.110 power plants could provide. 53:42.470 --> 53:47.650 A nuclear power plant usually has something 1 to 1.5 gigawatt maximum 53:47.650 --> 53:47.990 power. 53:48.210 --> 53:52.870 Certainly it's not needed all the time, but it's additional power that 53:52.870 --> 53:53.230 is needed. 53:54.270 --> 53:58.810 Now they will, as I said here, 20 kilowatt hours will last or will be 53:58.810 --> 54:00.730 sufficient for 100 kilometers driving. 54:01.530 --> 54:09.950 It means that normally during the day you don't need that much, so you 54:09.950 --> 54:14.250 will not need that much power over a longer period of time. 54:14.630 --> 54:19.290 But you have to look at that, like essentially we get quite a high 54:19.290 --> 54:23.130 demand for power at certain points in time. 54:23.390 --> 54:30.570 You could exactly compute the demand on looking at the average driving 54:30.570 --> 54:36.230 patterns of people, what is their mileage over a year, then you can 54:36.230 --> 54:41.850 calculate how much power they need for that distance that they drive. 54:41.930 --> 54:47.550 If they drive for 15,000 kilometers, then you need something like 30 54:47.550 --> 54:48.970 megawatt of power for that. 54:50.450 --> 54:51.850 Yeah, something like that. 54:52.290 --> 55:02.730 So we have at least the potential power that they can ask for, but if 55:02.730 --> 55:06.830 they can resupply, if they can feedback the power, then they could 55:06.830 --> 55:10.610 also supply power at the same rate. 55:12.310 --> 55:19.730 And if we would have three-phase charging at 10 or 11 kilowatt, then 55:19.730 --> 55:25.730 certainly we have that much extra power that we could offer to the 55:25.730 --> 55:26.110 system. 55:26.790 --> 55:32.130 And how long do they actually need to have the battery recharged, even 55:32.130 --> 55:33.230 if it's completely empty? 55:33.910 --> 55:37.430 If they do single-phase charging, it would take five to seven hours. 55:37.870 --> 55:43.110 If they do three-phase charging, it would take something about two 55:43.110 --> 55:43.530 hours. 55:43.630 --> 55:48.910 If you use not 16 ampere but 32 ampere, it would be just about an hour 55:48.910 --> 55:50.730 to recharge your battery completely. 55:52.690 --> 56:00.310 And it means that the rest of the time is available for moving back 56:00.310 --> 56:01.830 and forth this charging process. 56:02.390 --> 56:08.710 So you have a charging demand at some point in time, but the period 56:08.710 --> 56:10.510 where the car is sitting is much longer. 56:11.010 --> 56:17.250 So we have quite some flexibility on deciding when to use that extra 56:17.250 --> 56:18.650 load profile. 56:20.490 --> 56:25.770 So what we have here is quite a potential of high flexibility for load 56:25.770 --> 56:26.110 shifting. 56:27.150 --> 56:31.990 Certainly, we could also get quite a high peak load, and we have to 56:31.990 --> 56:35.970 control the charging processes in an intelligent way. 56:37.150 --> 56:41.730 If we would just plug in our vehicles and let them be charged, that's 56:41.730 --> 56:42.310 not reasonable. 56:43.030 --> 56:45.390 So we need more intelligent systems for that. 56:45.870 --> 56:48.550 That's one of the major things that people are working on at the 56:48.550 --> 56:49.510 moment in those projects. 56:50.310 --> 56:52.350 So let's look at what we would like to achieve. 56:52.430 --> 57:00.770 For example, if like this is some power diagram for a region of the 57:00.770 --> 57:06.490 network, actually having 100 households in some network segment, and 57:06.490 --> 57:13.190 it's a typical in black here the load profile in red or this is 57:13.190 --> 57:18.070 orange, that's photovoltaic power input idealized. 57:18.550 --> 57:23.490 And then assume we have uncontrolled electric vehicle charging. 57:24.350 --> 57:29.970 It would add additional power exactly at those times of the day, like 57:29.970 --> 57:35.590 in the evening times, where we have a peak demand anyway. 57:37.330 --> 57:42.130 So what we would like to do is shift that extra demand to the night 57:42.130 --> 57:42.590 hours. 57:44.610 --> 57:50.950 Now, this is one idea to move the demand into other times of the day, 57:51.610 --> 57:55.790 where there's less demand from other devices. 57:56.750 --> 58:01.730 And then you could say, I would like to drive on renewable energy. 58:02.370 --> 58:06.690 And then you would need the power to charge your battery exactly when 58:06.690 --> 58:09.290 photovoltaic systems are providing energy. 58:09.870 --> 58:15.170 Maybe you have your car at a parking lot at your working place, and 58:15.170 --> 58:17.510 then you could charge your battery there. 58:18.210 --> 58:21.430 And maybe when you are at home, you would like to reduce the peak 58:21.430 --> 58:27.090 demand of your house, and so feedback energy into the house. 58:27.830 --> 58:32.150 And it means that if you combine these things, you might end up in a 58:32.150 --> 58:35.750 load curve that is indicated in green here, which is very stable. 58:36.570 --> 58:42.330 Or maybe you can just generate a different load profile, which is more 58:42.330 --> 58:47.650 adequate to the needs of the distribution system that is behind. 58:47.850 --> 58:48.090 Yes. 59:22.390 --> 59:28.770 I don't talk, I don't say that this is the solution for providing 59:28.770 --> 59:31.950 energy at times where there is a shortage of energy. 59:33.190 --> 59:38.330 But we can contribute in some way to that, for that purpose. 59:38.570 --> 59:41.870 We can generate some contribution for that. 59:42.330 --> 59:47.170 But the major point is that here we have quite a significant extra 59:47.170 --> 59:48.750 load, extra demand. 59:48.750 --> 59:51.850 Like, these batteries have to be recharged. 59:53.070 --> 59:59.130 And we can shift that demand quite freely over quite some time. 59:59.210 --> 01:00:03.670 Like, we can either do it all in the evening hours, or we could shift 01:00:03.670 --> 01:00:06.110 it to some other time during the night. 01:00:07.150 --> 01:00:11.090 And this means, that means we have flexible demand. 01:00:12.330 --> 01:00:18.430 And we know about the flexibility at the moment when the driver plugs 01:00:18.430 --> 01:00:24.510 in his vehicle, or parks his vehicle, plugs in the cable, and tells me 01:00:24.510 --> 01:00:25.950 when he would like to leave again. 01:00:26.670 --> 01:00:29.790 Then I know how much flexibility I have. 01:00:30.510 --> 01:00:35.530 And it's just this flexibility of demand, which can be very beneficial 01:00:35.530 --> 01:00:36.310 for the system. 01:00:37.390 --> 01:00:41.110 And it's even sitting on the parking lot during the day, and again you 01:00:41.110 --> 01:00:42.170 have some flexibility. 01:00:42.170 --> 01:00:51.010 And if you have, like, it is a question whether it is economically, or 01:00:51.010 --> 01:00:55.510 whether it is a business case to provide power from your cars better. 01:00:56.550 --> 01:01:02.410 This depends on the cost you actually have to spend on that, like, on 01:01:02.410 --> 01:01:06.350 your, on getting the power. 01:01:06.450 --> 01:01:10.510 Like, you have an addition, maybe you have an additional cycle, but 01:01:10.510 --> 01:01:12.630 maybe you earn money from offering that power. 01:01:13.170 --> 01:01:19.870 Maybe your employer has 100 cars on the parking lot, and can combine 01:01:19.870 --> 01:01:24.330 all the power from that, from those batteries, and offer now this 01:01:24.330 --> 01:01:27.870 combined load on the market for balancing power. 01:01:29.510 --> 01:01:31.530 And this is an extra business case. 01:01:32.490 --> 01:01:37.450 And so it's very difficult in the moment to say, this is not feasible, 01:01:38.010 --> 01:01:39.350 or it is expensive. 01:01:39.550 --> 01:01:44.210 Like, the storage of power in the battery is not very efficient, 01:01:44.290 --> 01:01:48.570 because you can only have 80% efficiency if you look at storing and 01:01:48.570 --> 01:01:49.270 getting it back. 01:01:49.810 --> 01:01:50.790 So it's not efficient. 01:01:51.890 --> 01:01:53.230 This is not the story. 01:01:54.310 --> 01:01:59.790 The efficiency might increase, and you might be able to actually 01:01:59.790 --> 01:02:05.210 generate extra money just by offering this flexibility on some market. 01:02:05.690 --> 01:02:12.490 Or we might get beneficial effects just by shifting the charging time 01:02:12.490 --> 01:02:16.450 of the vehicles in an appropriate way. 01:02:18.030 --> 01:02:24.110 And if you are, assume you're operating a fleet of vehicles, you have 01:02:24.110 --> 01:02:27.410 some delivery service, and all your vehicles are electric vehicles, 01:02:28.030 --> 01:02:31.290 then you know that they will all be sitting on your parking lot during 01:02:31.290 --> 01:02:31.990 the night hours. 01:02:33.130 --> 01:02:39.130 Then you can control the demand for power in a very flexible way. 01:02:40.750 --> 01:02:42.690 Yeah, this is what you have to explore. 01:02:44.450 --> 01:02:50.290 Okay, now, this is the idea to have to get flexible adjustment of the 01:02:50.290 --> 01:02:53.850 load profile of those combined electric vehicles. 01:02:54.590 --> 01:03:00.270 And we are doing research on that in our energy smart home living lab 01:03:00.270 --> 01:03:05.290 at KIT, which is, like, this smart home is not looking like a smart 01:03:05.290 --> 01:03:09.630 home, because it's looking like some technical facility. 01:03:10.510 --> 01:03:13.710 And we were actually asked to let it look like that. 01:03:13.810 --> 01:03:16.110 We want it to look, like, much nicer. 01:03:16.710 --> 01:03:21.070 But they said, well, all the buildings in the environment look like 01:03:21.070 --> 01:03:22.830 that, so it has to look like that also. 01:03:23.710 --> 01:03:26.170 It's not sitting at that place in the moment. 01:03:26.270 --> 01:03:29.590 It had to be shifted to the right a little bit. 01:03:30.790 --> 01:03:36.350 And then we, well, what is inside that smart home? 01:03:36.690 --> 01:03:38.990 There is a two-bedroom apartment in there. 01:03:39.270 --> 01:03:42.290 I will give you more information on the individual components in the 01:03:42.290 --> 01:03:42.610 moment. 01:03:43.090 --> 01:03:46.670 So, there's also a car outside, or an Opel Meriva Meragio. 01:03:47.210 --> 01:03:48.710 There are living rooms. 01:03:49.370 --> 01:03:50.770 Actually, people can live there. 01:03:51.530 --> 01:03:56.110 We have some periods, some testing periods, where people are living 01:03:56.110 --> 01:03:57.730 there for six or eight weeks. 01:03:58.130 --> 01:04:02.810 Then we observe how they are actually using the components in that 01:04:02.810 --> 01:04:10.370 house, how they respond to our interfaces that we developed for energy 01:04:10.370 --> 01:04:11.330 control. 01:04:12.130 --> 01:04:15.290 And in that way, we get experience what we can actually do in such a 01:04:15.290 --> 01:04:16.090 smart home environment. 01:04:17.350 --> 01:04:20.650 But briefly, I should tell you something what the notion of a smart 01:04:20.650 --> 01:04:21.910 home actually is about. 01:04:22.410 --> 01:04:24.490 So, the notion of a smart home is quite old. 01:04:25.470 --> 01:04:29.330 And originally, it was just about home automation. 01:04:30.750 --> 01:04:35.190 So, remote switching of lights, of windows, of shutters, like window 01:04:35.190 --> 01:04:39.750 shutters, coffee machines, heating control, automatic light control, 01:04:39.870 --> 01:04:40.630 and things like that. 01:04:40.770 --> 01:04:42.490 Also, home security and safety. 01:04:43.030 --> 01:04:46.910 It's an important part in a smart home that you would like to have 01:04:46.910 --> 01:04:52.330 home security like only you, like when you get closer to the house, 01:04:52.590 --> 01:04:55.150 you're recognized and door gets open. 01:04:55.790 --> 01:04:58.590 If somebody else gets there, the door is shut. 01:04:58.890 --> 01:04:59.610 Things like that. 01:05:00.110 --> 01:05:02.770 Smoke alarm, smoke detection, things like that. 01:05:02.870 --> 01:05:03.830 And with assisted living. 01:05:05.290 --> 01:05:10.430 Assistance technologies for elderly people are also important. 01:05:10.870 --> 01:05:15.150 And you can observe the behavior of elderly people and then ask for 01:05:15.150 --> 01:05:18.690 assistance in a nice way. 01:05:19.890 --> 01:05:24.810 And then there is multimedia points so that you have music and 01:05:24.810 --> 01:05:31.110 pictures and all kinds of displays, like at a glance, in-house 01:05:31.110 --> 01:05:37.530 displays, like devices that you look at and you can control them by 01:05:37.530 --> 01:05:39.230 just saying something. 01:05:39.370 --> 01:05:43.570 You look at a device and say switch on and it switches on, only the 01:05:43.570 --> 01:05:44.510 device that you look at. 01:05:45.030 --> 01:05:47.690 Things like that are smart home environments. 01:05:48.910 --> 01:05:51.030 So it's a range of topics. 01:05:51.230 --> 01:05:52.470 There's no energy in there. 01:05:53.290 --> 01:05:57.130 And so when we talk about smart home, we say the energy smart home. 01:05:57.510 --> 01:05:59.430 So we are interested in the energy aspects. 01:06:00.130 --> 01:06:01.910 So that's what we add here. 01:06:02.390 --> 01:06:06.250 So comfort, security, and health are the traditional ingredients of a 01:06:06.250 --> 01:06:06.710 smart home. 01:06:07.090 --> 01:06:12.590 And we add here the energy aspect and how we can control the energy 01:06:13.910 --> 01:06:16.550 consumption and generation in such a house. 01:06:17.950 --> 01:06:22.090 So we have here all kinds of different scenarios for that. 01:06:22.530 --> 01:06:26.930 You can have a combination of smart metering technology and then some 01:06:26.930 --> 01:06:27.770 energy management. 01:06:28.090 --> 01:06:29.510 So control the devices. 01:06:31.030 --> 01:06:37.410 Quite a few devices and apps and so on are available for doing that, 01:06:37.970 --> 01:06:42.070 like based on smart metering technology to visualize the current power 01:06:42.070 --> 01:06:42.590 consumption. 01:06:44.030 --> 01:06:48.630 Then you could look at interdependencies between households and the 01:06:48.630 --> 01:06:49.410 distribution grid. 01:06:49.870 --> 01:06:53.590 How do you respond to bottlenecks in the distribution system? 01:06:54.650 --> 01:06:59.930 You could integrate decentralized power generation, micro combined 01:06:59.930 --> 01:07:07.130 heat and power plants, micro CHPs, and optimize their performance in 01:07:07.130 --> 01:07:07.550 households. 01:07:08.650 --> 01:07:12.750 You have photovoltaic systems on the rooftop, but usually you don't 01:07:12.750 --> 01:07:13.970 control them. 01:07:14.070 --> 01:07:19.250 But you have to, you might be able to control your systems depending 01:07:19.250 --> 01:07:21.050 on what they are generating. 01:07:22.090 --> 01:07:26.470 Then you have all kinds of appliances which might be controllable. 01:07:26.950 --> 01:07:30.910 There's again the washing machine that we talked about, the dryer, the 01:07:30.910 --> 01:07:32.790 deep freezer, the fridge, air conditioning. 01:07:34.310 --> 01:07:38.650 How to control a washing machine and dryer is a bit difficult, but 01:07:38.650 --> 01:07:40.990 deep freezer, fridge, and air conditioning is very simple. 01:07:42.050 --> 01:07:45.350 We have electric mobility, that's the major concern of that project. 01:07:46.050 --> 01:07:48.190 And then we have different optimization goals. 01:07:49.110 --> 01:07:53.570 The typical optimization goal is minimize energy consumption, increase 01:07:53.570 --> 01:07:54.570 energy efficiency. 01:07:55.650 --> 01:07:58.850 When you talk to people, they always say, okay, how much, how much can 01:07:58.850 --> 01:08:01.770 I save by using your smart controls? 01:08:03.750 --> 01:08:10.730 Another point is minimize carbon dioxide emissions, also important. 01:08:11.530 --> 01:08:15.530 Then maximize self-consumption of photovoltaic generation. 01:08:16.350 --> 01:08:20.790 That means shift your demand to times where you have photovoltaic 01:08:20.790 --> 01:08:25.930 power available, or store the photovoltaic power to times when you 01:08:25.930 --> 01:08:26.510 have the demand. 01:08:27.750 --> 01:08:32.130 So this is putting energy into those scenarios. 01:08:33.230 --> 01:08:37.310 Now let's look again at this KIT Energy Smart Home Lab, what we have 01:08:37.310 --> 01:08:37.570 there. 01:08:37.630 --> 01:08:43.190 So this is the layout, and we have intelligent appliances in the 01:08:43.190 --> 01:08:43.490 kitchen. 01:08:44.310 --> 01:08:51.130 We have a washing machine, we have a toaster, we have a stove, an 01:08:51.130 --> 01:08:57.850 oven, we have freezer, deep freezer, fridge, all kinds of things. 01:08:58.370 --> 01:09:00.850 These are the major things there. 01:09:01.170 --> 01:09:03.450 We have a combined heat and power plant. 01:09:04.070 --> 01:09:05.490 We have thermal storage. 01:09:06.170 --> 01:09:11.190 This thermal storage actually has also some heating device. 01:09:11.390 --> 01:09:13.190 So we have electrical heating there also. 01:09:13.310 --> 01:09:17.730 So we can switch between heating water in the directly with the 01:09:17.730 --> 01:09:23.450 electrical heater with eight kilowatt actually, or using the combined 01:09:23.450 --> 01:09:26.930 heat and power plant there we use gas to generate heat. 01:09:28.090 --> 01:09:31.570 So we can switch between gas and power. 01:09:32.450 --> 01:09:34.630 We have air conditioning. 01:09:35.950 --> 01:09:41.630 That's actually a phase shift material, some kind of salt, which is 01:09:41.630 --> 01:09:43.990 melting at, I think, 23 degrees. 01:09:44.490 --> 01:09:50.370 It's put in the seating, and if it's a temperature below 23 degrees, 01:09:50.430 --> 01:09:51.310 it's crystallized. 01:09:51.890 --> 01:09:55.490 And if temperature in the room gets above that, it's melting. 01:09:56.070 --> 01:10:00.830 This melting is using energy, and it's felt as a cooling effect. 01:10:01.030 --> 01:10:05.270 So we have a room that's automatically cooled by that melting of the 01:10:05.270 --> 01:10:08.350 crystals into this fluid state. 01:10:08.990 --> 01:10:13.030 And in this way, you can have air conditioning, which is just cooling 01:10:13.030 --> 01:10:14.130 down on demand. 01:10:14.850 --> 01:10:18.230 And you can then cool down the devices sometime during the day 01:10:18.230 --> 01:10:18.910 whenever you like. 01:10:19.530 --> 01:10:23.490 But the cooling effect is coming at the time when it's needed. 01:10:25.330 --> 01:10:27.170 Okay, that's the air conditioning. 01:10:27.450 --> 01:10:29.150 Then we have the photovoltaic cells. 01:10:29.270 --> 01:10:30.370 We have solar inverter. 01:10:30.370 --> 01:10:36.450 We actually also have a solar power generator or simulator. 01:10:36.590 --> 01:10:40.750 That's a device which the electrical engineers built for us, and they 01:10:40.750 --> 01:10:46.090 actually built a system which can replay the generation of 01:10:46.090 --> 01:10:47.070 photovoltaic power. 01:10:47.550 --> 01:10:51.150 So we have a certain profile, measured profile of electric power, of 01:10:51.150 --> 01:10:56.390 photovoltaic power, and then we can replay that at any time and 01:10:56.390 --> 01:11:00.670 simulate the effect of photovoltaic power generation. 01:11:01.410 --> 01:11:08.050 We can also scale it, like we have a five kilowatt peak system there. 01:11:08.350 --> 01:11:10.330 We can scale that to 20 kilowatt peak. 01:11:11.250 --> 01:11:16.190 And so we have a very flexible device there for actually using that as 01:11:16.190 --> 01:11:16.530 a lab. 01:11:17.190 --> 01:11:18.650 We have a smart metering system. 01:11:20.050 --> 01:11:25.610 We have another component, which I will tell you a little bit later. 01:11:26.050 --> 01:11:30.230 We have a charging station outside, which is actually capable of 01:11:30.230 --> 01:11:36.050 supporting bi-directional charging, like charging the battery or 01:11:36.050 --> 01:11:37.790 feeding back energy into the house. 01:11:39.230 --> 01:11:42.750 And the car is supporting that only, that also it is the only car that 01:11:42.750 --> 01:11:49.410 is actually having that capability, designed by Opel, together with 01:11:49.410 --> 01:11:49.630 us. 01:11:50.210 --> 01:11:53.350 And then we have the energy management panels. 01:11:54.150 --> 01:11:58.630 These energy management panels are actually showing the current 01:11:58.630 --> 01:11:59.890 situation of the power grid. 01:12:00.030 --> 01:12:05.130 I will show you in a moment the kinds of devices that we see there. 01:12:05.390 --> 01:12:07.670 Oops, I didn't want that. 01:12:08.910 --> 01:12:10.990 This is really annoying. 01:12:11.770 --> 01:12:13.530 That is Windows 8. 01:12:14.210 --> 01:12:19.670 Now, if you touch the side of the screen, you get extra functionality. 01:12:20.370 --> 01:12:23.430 So this is some kind of information. 01:12:23.550 --> 01:12:26.170 I will show it to you in larger size in a moment. 01:12:26.630 --> 01:12:31.130 So we can visualize energy consumption, energy usage, and we can 01:12:31.130 --> 01:12:32.530 discover user preferences. 01:12:33.090 --> 01:12:34.890 Then we come to your question. 01:12:35.590 --> 01:12:40.790 And then we have an energy management system, which is observing 01:12:40.790 --> 01:12:45.750 what's happening in the house, which is collecting all the information 01:12:45.750 --> 01:12:50.350 on the user preferences that are detecting using the energy management 01:12:50.350 --> 01:12:50.770 panels. 01:12:51.570 --> 01:12:55.710 And then this information is used to control the electric and thermal 01:12:55.710 --> 01:12:59.250 consumers and providers in the house in an optimized way. 01:13:00.230 --> 01:13:02.710 So we have an optimization algorithm behind that. 01:13:03.310 --> 01:13:07.790 We also know what the expected usage of devices will be during the 01:13:07.790 --> 01:13:10.850 next day, or during the remainder of this day. 01:13:11.430 --> 01:13:16.090 And then we can optimally plan the use of devices. 01:13:16.900 --> 01:13:22.150 So this is showing you that all these devices are connected to some 01:13:22.150 --> 01:13:23.470 server system. 01:13:23.990 --> 01:13:27.690 Here's again what we're interested in, the interplay between humans, 01:13:28.210 --> 01:13:30.250 the home, and the intelligent devices. 01:13:31.410 --> 01:13:35.210 And this is just summarizing again all the different components there. 01:13:35.770 --> 01:13:38.770 And we have one more component which is of interest, which is a four 01:13:38.770 --> 01:13:39.750 -quadrant amplifier. 01:13:40.310 --> 01:13:45.390 This allows to actually generate arbitrary situations in the house. 01:13:45.650 --> 01:13:50.030 So it means we are connected to the medium-voltage system, which is 01:13:50.030 --> 01:13:50.630 very stable. 01:13:51.270 --> 01:13:57.350 But we would like to see in which way we can respond to a critical 01:13:57.350 --> 01:14:01.030 situation, the grid, which we never see in the system here. 01:14:01.570 --> 01:14:04.470 So we can generate that using the four-quadrant amplifier. 01:14:05.210 --> 01:14:10.910 And then we can see how we can actually control the devices in the 01:14:10.910 --> 01:14:15.490 house, provide certain energy system services, such that we can 01:14:15.490 --> 01:14:16.370 improve the system. 01:14:17.170 --> 01:14:20.910 For that we need this four-quadrant amplifier, which allows to use 01:14:20.910 --> 01:14:24.610 actually that house in some kind of hardware-in-the-loop simulation 01:14:24.610 --> 01:14:25.110 system. 01:14:26.790 --> 01:14:31.070 Okay, this is the electric vehicle part. 01:14:31.170 --> 01:14:34.570 This is the Opel Meriva Meragio that is sitting outside there. 01:14:35.370 --> 01:14:41.290 We also promised to get a similar car from Daimler, which we never 01:14:41.290 --> 01:14:41.650 did. 01:14:42.510 --> 01:14:47.190 And there are some more electric vehicles running around in the field 01:14:47.190 --> 01:14:53.450 test, which sometimes also visited our smart home, but very seldom. 01:14:54.890 --> 01:14:59.910 And the focus of our approach is that we would like to develop a 01:14:59.910 --> 01:15:01.330 demand -side load management. 01:15:02.110 --> 01:15:07.610 These appliances are scheduled based on some time-varying signal, 01:15:07.770 --> 01:15:14.730 price signal, but also with respect to some power constraint signals. 01:15:15.690 --> 01:15:20.570 And the idea is that we would like to get an increasing... or the 01:15:20.570 --> 01:15:24.790 challenge certainly is that we have an increasing imbalance of 01:15:24.790 --> 01:15:27.050 generator and consumer, as you know, meanwhile. 01:15:27.590 --> 01:15:29.990 So we would like to deal with that. 01:15:30.730 --> 01:15:35.290 These are the degrees of freedom that we have in our devices. 01:15:36.170 --> 01:15:39.030 They are different, and so we can classify our appliances. 01:15:39.810 --> 01:15:44.630 So some of them have a very poor degree of freedom. 01:15:45.690 --> 01:15:48.550 The TV set is not really controllable. 01:15:50.490 --> 01:15:55.050 You will not switch off the TV set just because the power is 01:15:55.050 --> 01:15:55.450 expensive. 01:15:55.970 --> 01:15:57.470 You would like to watch the news. 01:15:58.330 --> 01:16:03.050 And at lunchtime you would like to use the oven or the stove for 01:16:03.050 --> 01:16:03.390 cooking. 01:16:04.770 --> 01:16:09.890 And then there are some which are more reschedulable, like the deep 01:16:09.890 --> 01:16:11.250 freezer, the electric heating. 01:16:12.570 --> 01:16:16.650 The washing machine, the dishwasher can only be used if they have been 01:16:16.650 --> 01:16:22.230 specified that they should run sometime in the next five hours. 01:16:22.550 --> 01:16:24.850 Then you can use that information. 01:16:25.790 --> 01:16:29.810 And so we can classify these household appliances in different ways. 01:16:30.490 --> 01:16:35.970 Some are controllable, like a deep freezer, the heating, air 01:16:35.970 --> 01:16:38.590 conditioning, warm water boiler, and so on. 01:16:41.030 --> 01:16:46.210 For some, like a timed service, for certain times of time periods, 01:16:46.770 --> 01:16:49.330 dishwasher, washing machine, and dryer might be controllable. 01:16:49.810 --> 01:16:51.970 And some are only observable. 01:16:52.670 --> 01:16:56.030 You know at which times they will usually be used. 01:16:56.830 --> 01:17:01.570 But so it's important to know that they will generate a certain load 01:17:01.570 --> 01:17:04.350 profile, but they are not controllable. 01:17:06.470 --> 01:17:09.790 And so what we have here is this 24-hour bright signal. 01:17:10.910 --> 01:17:18.290 And then we can, like this is the time-dependent signal, communicated 01:17:18.290 --> 01:17:21.650 by the energy provider in our smart home, actually generated locally, 01:17:21.890 --> 01:17:23.950 independent of the energy provider. 01:17:24.790 --> 01:17:31.770 And then we can try to actually look at the current schedule of our 01:17:31.770 --> 01:17:32.250 devices. 01:17:32.590 --> 01:17:36.530 So here we have a toaster, a bread maker, a deep freezer, like the 01:17:36.530 --> 01:17:39.990 regular iterative profile of a deep freezer. 01:17:40.210 --> 01:17:42.290 We have the stove, we have the dishwasher. 01:17:42.930 --> 01:17:48.490 Then we see the time or the time-varying price signal, and we could 01:17:48.490 --> 01:17:50.210 try to optimize the schedule. 01:17:50.450 --> 01:17:53.990 That's one idea, just by looking at the price signal. 01:17:54.370 --> 01:17:57.690 But there's always the problem of getting evident facts, which I 01:17:57.690 --> 01:17:58.430 mentioned before. 01:18:00.530 --> 01:18:02.450 Why did that happen? 01:18:03.070 --> 01:18:16.490 Because I forgot to put in... I thought I had taken the... 01:18:16.490 --> 01:18:19.890 Yes, I have it here. 01:18:20.770 --> 01:18:23.390 Oops, I did not put it in. 01:18:24.470 --> 01:18:25.130 It's there. 01:18:27.550 --> 01:18:29.290 So, just a moment. 01:18:31.190 --> 01:18:32.490 Now the system is happy again. 01:18:34.750 --> 01:18:37.270 Okay, getting energy, getting power. 01:18:38.270 --> 01:18:42.510 Okay, so then we could just try to move that to better times. 01:18:42.510 --> 01:18:45.790 And so this... but this typical schedule is something which the system 01:18:45.790 --> 01:18:48.390 could learn by observing the daily schedule. 01:18:48.850 --> 01:18:50.570 Then you can use that appropriately. 01:18:51.550 --> 01:18:55.130 Then you could optimize the appliances appropriately. 01:18:56.190 --> 01:19:02.730 And so, for example, you have certain objectives to minimize the cost 01:19:04.630 --> 01:19:08.310 and minimize the current power limit value violation. 01:19:08.810 --> 01:19:13.390 So you have certain power limitation signals in addition to these 01:19:13.390 --> 01:19:14.170 price signals. 01:19:14.910 --> 01:19:18.890 And you could try to comply with both objectives. 01:19:19.650 --> 01:19:24.950 And then you would like to adapt the system such that, for example, 01:19:25.450 --> 01:19:30.850 the bread machine is starting earlier because you don't like to use 01:19:30.850 --> 01:19:32.650 power when the price gets up. 01:19:32.650 --> 01:19:36.350 But you would like to do that only at the time when the price is low. 01:19:37.230 --> 01:19:45.230 You might want to move the dishwasher to a later time or because of 01:19:45.230 --> 01:19:48.790 the price signals or in the evening hours, washing machine and 01:19:48.790 --> 01:19:50.210 dishwasher to a later time. 01:19:51.750 --> 01:19:58.330 You cannot use or modify the stove usage, but you could use power from 01:19:58.330 --> 01:19:58.790 the battery. 01:19:59.670 --> 01:20:03.690 And so these are options that you could use in optimizing the power 01:20:03.690 --> 01:20:04.710 schedule for a day. 01:20:05.010 --> 01:20:08.390 So these are things that the demand management could do. 01:20:09.190 --> 01:20:11.090 You could then recharge the battery at night. 01:20:11.530 --> 01:20:16.410 This is a simple scenario that we looked at, essentially also looked 01:20:16.410 --> 01:20:16.830 at already. 01:20:18.230 --> 01:20:22.190 So this is, again, showing you the interior space in the house. 01:20:22.790 --> 01:20:28.190 And now we come to the interesting ingredient, the energy management 01:20:28.190 --> 01:20:28.490 panel. 01:20:28.570 --> 01:20:30.690 What is the task of the energy management panel? 01:20:31.330 --> 01:20:36.450 We would like to make the information on energy consumption very 01:20:36.450 --> 01:20:37.810 transparent to the user. 01:20:39.850 --> 01:20:40.770 Very simple way. 01:20:40.970 --> 01:20:45.670 So this is not needing sophisticated devices, but this is just a web 01:20:45.670 --> 01:20:51.390 service which can be displayed on any screen, preferably on a touch 01:20:51.390 --> 01:20:53.450 screen, maybe on your smartphone. 01:20:54.150 --> 01:20:54.670 No problem. 01:20:54.790 --> 01:20:55.830 Can be used on any smartphone. 01:20:56.890 --> 01:21:01.250 And then we would also like to use those displays to discover and 01:21:01.250 --> 01:21:03.670 specify degrees of freedom for energy consumption. 01:21:04.370 --> 01:21:06.710 At what time can we actually use the washing machine? 01:21:07.070 --> 01:21:07.890 That was your question. 01:21:09.810 --> 01:21:16.930 So we would like to provide information on what's currently the 01:21:16.930 --> 01:21:20.630 situation of the energy system, not just currently, but also in the 01:21:20.630 --> 01:21:21.110 past. 01:21:21.630 --> 01:21:26.250 So here you see the power consumption for the last 24 hours. 01:21:26.990 --> 01:21:29.270 You see the time-varying price signal. 01:21:29.910 --> 01:21:30.930 The current price. 01:21:31.190 --> 01:21:33.690 You see the power consumption of the household. 01:21:34.570 --> 01:21:40.790 The power that you actually get here from the utility. 01:21:42.390 --> 01:21:49.350 Then the combined heat and power plant is somehow this looking a bit 01:21:49.350 --> 01:21:54.030 strange, but this is a real screenshot, so it should be reasonable. 01:21:54.190 --> 01:21:58.190 It's showing the power that you use inside the house for the different 01:21:58.190 --> 01:22:02.370 components, how much you get from the photovoltaic system, and so on. 01:22:02.970 --> 01:22:09.750 You can click on the photovoltaic system and you get on the device in 01:22:09.750 --> 01:22:16.490 a different menu here, and you get information on the power generation 01:22:16.490 --> 01:22:18.510 history of that photovoltaic system. 01:22:18.610 --> 01:22:23.370 This is just showing you the situation, what kind of things have been 01:22:23.370 --> 01:22:24.910 occurring in the house. 01:22:26.090 --> 01:22:31.390 And then you can also click on a layout of, if you have the household 01:22:31.950 --> 01:22:36.130 overview, you click on the kitchen area and you get information on all 01:22:36.130 --> 01:22:40.850 the components there, and those that are running are indicated in 01:22:40.850 --> 01:22:43.790 yellow, the other ones are currently not running. 01:22:44.930 --> 01:22:49.790 And then you can click on any of those devices and then you get even 01:22:49.790 --> 01:22:50.570 more information. 01:22:51.590 --> 01:23:02.130 You get information, for example, on the tumble dryer and there you 01:23:02.130 --> 01:23:06.510 can indicate the degree of freedom, the time interval where it should 01:23:06.510 --> 01:23:07.050 be used. 01:23:07.270 --> 01:23:12.470 So, assume you switch on, at the device, you switch on the tumble 01:23:12.470 --> 01:23:16.590 dryer, you fill it, switch it on, no, the tumble dryer, you put in all 01:23:16.590 --> 01:23:22.370 the laundry, the tumble dryer should dry your laundry, and then you 01:23:22.370 --> 01:23:30.090 say, okay, now I specify the period of time that I allow for carrying 01:23:30.090 --> 01:23:31.470 out the drying process. 01:23:33.290 --> 01:23:38.630 It need not be ready immediately or done immediately, the system tells 01:23:38.630 --> 01:23:41.810 you at what time it will start the drying process. 01:23:42.310 --> 01:23:45.930 If you don't like that, you can immediately switch it on. 01:23:48.010 --> 01:23:48.590 You have a question? 01:23:57.260 --> 01:23:57.680 Yes? 01:24:21.700 --> 01:24:24.140 But if you don't like that, then you don't do that. 01:24:24.700 --> 01:24:31.120 But if you are willing to say, okay, like, you have the washing 01:24:31.120 --> 01:24:35.720 machine in your basement, there, you don't mind whether it smells a 01:24:35.720 --> 01:24:37.420 little bit, and even if it does not... 01:24:39.660 --> 01:24:40.140 Yes? 01:24:43.040 --> 01:24:43.400 Yes? 01:24:46.640 --> 01:24:47.880 Well, I don't know. 01:24:49.040 --> 01:24:50.840 I don't see that in my washing machine. 01:24:51.940 --> 01:24:52.420 Yeah. 01:24:53.780 --> 01:24:58.080 But this is just showing the idea behind that. 01:24:58.680 --> 01:25:02.880 You just operate the machine, you put in, for example, in the 01:25:02.880 --> 01:25:07.000 dishwasher, you put in all the dishes, you say, I would like to have 01:25:07.000 --> 01:25:10.180 the dishes washed when I come back home. 01:25:11.000 --> 01:25:16.780 So I offer a certain time period, which is the time period where it 01:25:16.780 --> 01:25:17.460 should be used. 01:25:18.320 --> 01:25:19.860 And I don't want to start it now. 01:25:20.100 --> 01:25:22.740 I'm not at home to start it two hours later. 01:25:23.160 --> 01:25:27.780 But the system should choose the most appropriate time for doing that 01:25:27.780 --> 01:25:28.100 washing. 01:25:29.160 --> 01:25:31.060 And then the system can do that. 01:25:31.140 --> 01:25:34.280 And it will tell you, actually, when it, depending on the current 01:25:34.280 --> 01:25:35.640 information, when it will do that. 01:25:36.360 --> 01:25:38.640 And if you don't like that, you just switch it on immediately. 01:25:38.840 --> 01:25:45.200 Or maybe you could extend that and offer a possibility to set a 01:25:45.200 --> 01:25:47.280 specific time when it should start. 01:25:47.740 --> 01:25:49.740 With an easy extension of that. 01:25:50.800 --> 01:25:50.920 Yeah? 01:25:51.040 --> 01:25:56.080 But if you say, I don't worry, it should be done when I come back, 01:25:57.440 --> 01:25:59.100 then you just specify that time. 01:25:59.740 --> 01:26:01.600 You can easily modify that. 01:26:02.140 --> 01:26:06.800 And in this way, you have a possibility to specify your preferences 01:26:06.800 --> 01:26:08.440 for using the devices. 01:26:08.900 --> 01:26:13.480 If you are very restrictive, you don't offer flexibility. 01:26:14.780 --> 01:26:19.060 But if you say, I don't mind, I leave home in the morning and I'm 01:26:19.060 --> 01:26:23.800 coming back sometime in the afternoon, then I have that many hours for 01:26:23.800 --> 01:26:26.460 the energy management system to do with the system what it's like. 01:26:27.060 --> 01:26:29.160 Or to do with the devices what it's like. 01:26:29.480 --> 01:26:30.420 Or what it likes. 01:26:31.060 --> 01:26:32.480 What is appropriate. 01:26:33.220 --> 01:26:38.860 And then, you have the best possible control of those devices. 01:26:39.040 --> 01:26:41.380 But only complying with your preferences. 01:26:42.280 --> 01:26:42.420 Yeah? 01:26:42.440 --> 01:26:44.000 That's our philosophy behind it. 01:26:44.700 --> 01:26:48.540 We don't want to restrict the freedom of the user to do with the 01:26:48.540 --> 01:26:49.620 devices what they like. 01:26:50.380 --> 01:26:55.080 But if they are willing to offer flexibility, we can utilize that. 01:26:56.000 --> 01:26:56.400 Yeah? 01:26:56.500 --> 01:27:00.120 And that's what we will look at or continue to look at next week. 01:27:00.600 --> 01:27:01.220 Thank you. 01:27:01.680 --> 01:27:02.400 That's it for today.