WEBVTT 00:00.000 --> 00:06.000 Well, I come everybody. 00:06.000 --> 00:08.000 We're going to talk about Sun Peak. 00:08.000 --> 00:09.000 My name is Philip. 00:09.000 --> 00:10.000 This is Marna. 00:10.000 --> 00:13.000 And Sun Peak is about solar thermal plants. 00:13.000 --> 00:15.000 Solar thermal plants. 00:15.000 --> 00:18.000 It's an open source software for performance monitoring 00:18.000 --> 00:20.000 and assessment of such plants. 00:20.000 --> 00:22.000 So why even bother about this? 00:22.000 --> 00:24.000 Well, very simple, right? 00:24.000 --> 00:26.000 Everybody vaguely interested in the future 00:26.000 --> 00:29.000 the future of human kind probably knows this. 00:29.000 --> 00:32.000 Most of our greenhouse gas emissions come from energy. 00:32.000 --> 00:35.000 So we should bother about renewable energy. 00:35.000 --> 00:37.000 What is energy? 00:37.000 --> 00:40.000 Is it just what comes out of the circuit? 00:40.000 --> 00:43.000 Also, but not alone. 00:43.000 --> 00:46.000 Look at about half of our finite energy consumption 00:46.000 --> 00:48.000 is heating and cooling. 00:48.000 --> 00:51.000 So we can do part of that with heat pumps and stuff. 00:51.000 --> 00:52.000 It's good. 00:52.000 --> 00:54.000 Can electrify somethings. 00:54.000 --> 00:55.000 Still, it's half. 00:55.000 --> 00:56.000 Heat is half. 00:56.000 --> 00:59.000 And almost everything of that just 10% to 11%. 00:59.000 --> 01:02.000 Maybe 11.5 now. 01:02.000 --> 01:03.000 12%. 01:03.000 --> 01:06.000 But almost everything is fossil. 01:06.000 --> 01:08.000 So we have a very long way to go. 01:08.000 --> 01:11.000 Every is talking about increasing shares of renewables. 01:11.000 --> 01:12.000 Are we doing good? 01:12.000 --> 01:15.000 Yeah, we are slightly increasing in percent. 01:15.000 --> 01:17.000 But in absolute numbers, we are not. 01:17.000 --> 01:19.000 Look at this 10-year comparison. 01:19.000 --> 01:21.000 In absolute numbers, we are going up. 01:21.000 --> 01:24.000 We should be going down to zero as quickly as possible. 01:24.000 --> 01:27.000 So we have a long way to go. 01:27.000 --> 01:30.000 And there is big potential for lots of technologies. 01:30.000 --> 01:32.000 And solar thermal is one of them. 01:32.000 --> 01:36.000 Solar thermal has very good projections on the long run. 01:36.000 --> 01:39.000 So what is solar thermal? 01:39.000 --> 01:42.000 I guess most of you know PV for the world types by now. 01:42.000 --> 01:43.000 What is it? 01:43.000 --> 01:47.000 It's simply put devices that turn sunlight into electricity. 01:47.000 --> 01:51.000 This is a, maybe a simplified. 01:51.000 --> 01:53.000 But you understand the point. 01:53.000 --> 01:55.000 Solar thermal is what we talk about now. 01:55.000 --> 01:56.000 And it's very different. 01:56.000 --> 01:58.000 It's a very different technology. 01:58.000 --> 02:02.000 It's basically water or some other heat transfer fluid circulating in some collectors. 02:02.000 --> 02:06.000 And thanks to some highly efficient materials and coatings. 02:06.000 --> 02:09.000 It enters cold and it gets out hot. 02:09.000 --> 02:12.000 That's how it works. 02:12.000 --> 02:14.000 It's highly efficient. 02:14.000 --> 02:18.000 It means per square meter, it's about three times as efficient as PV. 02:18.000 --> 02:20.000 Of course, it produces heat. 02:20.000 --> 02:21.000 Now for electricity. 02:21.000 --> 02:23.000 So it's something different. 02:23.000 --> 02:29.000 Producers heat means it's only useful when it produces the temperature level that you need that you want. 02:29.000 --> 02:36.000 So we have a lot of different technologies for different temperature levels, including concentrating, including vacuum, etc. 02:36.000 --> 02:42.000 So you can tell we need a lot of different physical models to describe how all these technologies work. 02:43.000 --> 02:46.000 And this is basically what Sandpick is about. 02:46.000 --> 02:48.000 We focus on large scale systems. 02:48.000 --> 02:50.000 They look more or less like these. 02:50.000 --> 02:55.000 You can tell that there's a lot of different technologies involved. 02:55.000 --> 03:01.000 You can also tell from this picture that Sandpick focuses basically on a B2B level. 03:01.000 --> 03:03.000 So we don't deal with end customers primarily. 03:03.000 --> 03:05.000 It's B2B. 03:05.000 --> 03:08.000 And what we do with these, with these plants. 03:08.000 --> 03:10.000 We are not into designing them. 03:10.000 --> 03:12.000 We're not into installing them. 03:12.000 --> 03:19.000 We come into play when these plants are all out there and being commissioned and start the operation and start collecting data. 03:19.000 --> 03:22.000 Start collecting sensor data. 03:22.000 --> 03:27.000 And then with that sensor data, we try to answer basically two important questions. 03:27.000 --> 03:30.000 One is a verification question. 03:30.000 --> 03:33.000 Does the plant perform as expected? 03:33.000 --> 03:36.000 Well, it performs as it performs, like we measure it with heat meters. 03:36.000 --> 03:37.000 But what is expected? 03:37.000 --> 03:40.000 That's difficult because temperatures are different. 03:40.000 --> 03:41.000 The technologies are different. 03:41.000 --> 03:43.000 Weather is different, et cetera, et cetera. 03:43.000 --> 03:48.000 So the calculating the expected part is the heart part of this question. 03:48.000 --> 03:51.000 And the second is a monitoring question. 03:51.000 --> 03:54.000 Are there any changes over time? 03:54.000 --> 03:59.000 So again, we have very much into the operational phase of these plants. 03:59.000 --> 04:00.000 Why? 04:00.000 --> 04:03.000 Because on the design phase, there is love knowledge. 04:03.000 --> 04:06.000 It's quite well known by now how to build these plants. 04:06.000 --> 04:09.000 There is a lot of well-known and established standards. 04:09.000 --> 04:14.000 On the operational side, there is just one technical standard, which emerged two years ago. 04:14.000 --> 04:15.000 So it's very recent. 04:15.000 --> 04:22.000 And we're basically implementing methods from that ISO standard into our software. 04:22.000 --> 04:28.000 Because we believe that open source software is much more useful than PDF documents. 04:28.000 --> 04:35.000 And we're also active in the technical committee, which developed this ISO 24194. 04:35.000 --> 04:41.000 So to give you a few results of what it roughly does. 04:41.000 --> 04:46.000 Basically, some data filtering. 04:46.000 --> 04:48.000 And then we compare what is measured. 04:48.000 --> 04:51.000 What is measured at the plant level with the target. 04:51.000 --> 04:57.000 So with what some pick, platilates is the expected outcome. 04:57.000 --> 05:02.000 And yeah, if we are above what we expect, then we're doing good. 05:02.000 --> 05:04.000 So we do this for some period of time. 05:04.000 --> 05:07.000 And the measure is more than the target, more than the expected, 05:07.000 --> 05:09.000 than the check is verified. 05:09.000 --> 05:10.000 The check is passed. 05:10.000 --> 05:15.000 The second application of this can be to apply this check over time. 05:15.000 --> 05:19.000 And calculate this ratio of measured to expected. 05:19.000 --> 05:22.000 And then you see when the performance started degrade. 05:22.000 --> 05:27.000 And this is with all influencing factors already factored in already calculated. 05:27.000 --> 05:31.000 So this is not because of differences in weather or radiation and temperature. 05:31.000 --> 05:34.000 That's already factored in in our models. 05:34.000 --> 05:36.000 So a decrease is a decrease. 05:36.000 --> 05:40.000 So in that case, you can tell that something is going wrong at the plant. 05:40.000 --> 05:43.000 Either there is something that doesn't work. 05:43.000 --> 05:46.000 Or like in this simple case, it's just time to clean the collectors. 05:46.000 --> 05:49.000 But it's also very useful to know when it's time to clean the collectors. 05:49.000 --> 05:52.000 Because that costs money, right? 05:52.000 --> 05:56.000 Okay, and before I hand over to you, Marna, 05:57.000 --> 05:59.000 like a proud image that we submitted. 05:59.000 --> 06:01.000 I'll just paper just last week. 06:01.000 --> 06:04.000 So you can find a lot more about more research methods. 06:04.000 --> 06:08.000 That reactive in this was just a very simplified version of all. 06:08.000 --> 06:12.000 And we also share open data and more research papers. 06:12.000 --> 06:14.000 Or in our eyes in order community. 06:14.000 --> 06:16.000 Talking about some communities. 06:16.000 --> 06:18.000 I'll hand over to you. 06:22.000 --> 06:25.000 Marna comes from Scotland and then will explain everything in proper English. 06:26.000 --> 06:27.000 Sorry, Marna. 06:31.000 --> 06:36.000 Right, yes, somewhat unconventionally putting the community slide in the middle of the presentation. 06:36.000 --> 06:38.000 We should mention our funders. 06:38.000 --> 06:41.000 We're funded at the moment mostly by research grants. 06:41.000 --> 06:47.000 We're trying to change that from both EU and Austrian national level as well as some industry foundations. 06:48.000 --> 06:56.000 And we have a small but growing community of users and and. 06:56.000 --> 06:59.000 We say expertise contributors. 06:59.000 --> 07:03.000 We are really trying to grow our also software contributors. 07:03.000 --> 07:06.000 It's something we would very much like more of. 07:06.000 --> 07:12.000 From mostly academia at the moment with one or two industry users. 07:17.000 --> 07:27.000 Yes, so the project was initiated by ourselves as a research institute and then for commercial partners who who operate in the in the space. 07:27.000 --> 07:33.000 So we've been very close to the actual industry for the entire duration of the project. 07:33.000 --> 07:35.000 Right. 07:35.000 --> 07:39.000 Going to a little bit more detail about how. 07:39.000 --> 07:42.000 How this actually works so. 07:42.000 --> 07:50.000 The basic workflow in the software is that you add a solar thermal plant object. 07:50.000 --> 07:54.000 And then you have to describe the properties of that plant. 07:54.000 --> 07:59.000 So both the overall size of the field position. 07:59.000 --> 08:04.000 Various things like that as well as the specific model of collective that you're using. 08:05.000 --> 08:09.000 We include a fairly small database of those at the moment. 08:09.000 --> 08:13.000 We are actively working to integrate it with the. 08:13.000 --> 08:19.000 Sort of European industry standard database of certified solar thermal collectors. 08:19.000 --> 08:27.000 The reason picking a particular collect the model is very important is that that is how we calculate the expected performance. 08:27.000 --> 08:31.000 So that's a key step in every collective forms. 08:31.000 --> 08:35.000 Very significantly differently. 08:35.000 --> 08:39.000 We have a process called sensor mapping. 08:39.000 --> 08:45.000 Basically we try to be as open as possible in what input data we accept. 08:45.000 --> 08:50.000 Because every scatter system, every logging system outputs its own. 08:50.000 --> 08:55.000 Fun different data format with its own channel names. 08:55.000 --> 09:00.000 So rather than requiring our users to reformat all of their data in order to work with our software. 09:00.000 --> 09:08.000 We'll try and take as many different data formats as we feasibly can or at least as many different data layouts as we feasibly can. 09:08.000 --> 09:12.000 So that means we have defined. 09:12.000 --> 09:18.000 Roles is probably the best word for data channels within the software. 09:18.000 --> 09:29.000 And the user has did a one-off process to manually say this channel in my data with this arbitrary name that my control system engineer selected maps to this role. 09:30.000 --> 09:43.000 Upload data sets something called the safety factor, which is something that comes from the standard and is a slightly hand-wave deal with all of the sensor errors and losses from pipes and things factor. 09:43.000 --> 09:54.000 We would very much like to reduce how large that factor has to be and make it more precise, but at the moment it's kind of a requirement. 09:54.000 --> 10:00.000 I was going to go into a little bit more technical detail. 10:00.000 --> 10:04.000 I think for time I will jump over this slide. 10:04.000 --> 10:10.000 And yes, then I was going to go less technical detail and I'm going to architecture. 10:10.000 --> 10:23.000 But I just want to highlight the fact that we took an approach which has led to some slightly interesting packaging decisions and a little bit of complexity on our side. 10:23.000 --> 10:31.000 Of basically we want to support quite a wide range of users from what we deemed technical experts. 10:31.000 --> 10:44.000 So other researchers, other software engineers, people who actually want to work with our call libraries, directly either to integrate them into the software or to integrate them with a web API. 10:45.000 --> 11:07.000 And we also want to support general users from the plant operator companies who are typically not software engineers are typically not IT people in quite a lot of cases are users have an IT staff of zero, which means we need to be quite friendly to them. 11:07.000 --> 11:15.000 So we have a JavaScript front end, which is built as a separate component. 11:15.000 --> 11:20.000 And then our entire back end is in Python, I don't remember mentioned that yet. 11:20.000 --> 11:24.000 And we provide different packaging formats for our different users. 11:24.000 --> 11:28.000 So for the general user case, get it running quickly. 11:28.000 --> 11:34.000 We provide a set of Docker images and a Docker compose file. 11:35.000 --> 11:45.000 And then we also provide a Python package, which is on pipeline if anyone wants to use that. 11:45.000 --> 11:48.000 Yes, again different installation options. 11:48.000 --> 11:53.000 I'll just mention them for completeness on the slide. 11:53.000 --> 11:58.000 We also bundle a set of open data with the application. 11:58.000 --> 12:11.000 So one of our initial contributing cooperating partners contributed some data from one of their plants, which is normally very hard to get people don't like to share their operating data. 12:11.000 --> 12:22.000 So if you install the full application or you install the Python package with the demo extra you get a set of data to play around without having to actually set up your own plan to talk. 12:23.000 --> 12:28.000 And that is also available on we have a public demo server. 12:28.000 --> 12:35.000 Please be aware it is a public demo server any day to upload there is public. 12:35.000 --> 12:43.000 I have big plans in the back of my head to build a sandbox integrated with the documentation that that hasn't happened yet. 12:44.000 --> 12:50.000 Yes, just the last information is available. We have Doc site. 12:50.000 --> 12:52.000 We're on get lab. 12:52.000 --> 12:57.000 We want to join us and yes, please join us. 12:57.000 --> 13:03.000 If anyone is interested as joining as a contributor we will welcome anyone. 13:03.000 --> 13:06.000 I think we have time for questions. 13:06.000 --> 13:08.000 Okay. 13:13.000 --> 13:28.000 So the question was will it work on household sized systems and the answer is no. 13:28.000 --> 13:35.000 Mostly because they tend not to have the right instrumentation. 13:35.000 --> 13:40.000 Yes, and the answer is also yes. 13:40.000 --> 13:47.000 There are more methods that don't even require a lot of instrumentation. 13:47.000 --> 13:52.000 You set the data for methods that we did not present today. 13:52.000 --> 13:55.000 But we not targeting small household. 13:55.000 --> 13:58.000 Why? Because the market is decreasing a lot in small households. 13:58.000 --> 14:05.000 So we deliberately chose large installations because those are used more in industry and district heating applications. 14:05.000 --> 14:08.000 But in theory, a lot of this would also work. 14:08.000 --> 14:11.000 No one just is. 14:23.000 --> 14:26.000 We basically take everything that we can get. 14:26.000 --> 14:30.000 We take all temperature sensors. 14:30.000 --> 14:35.000 And the system for collector fields typically have at least one inlet one outlet temperature sensor. 14:35.000 --> 14:39.000 Sometimes they have one outlet temperature sensor per row. 14:39.000 --> 14:43.000 Because we are different so we need to control volume flows. 14:43.000 --> 14:49.000 We take thermal power output or volume flow or whatever we have. 14:49.000 --> 14:51.000 It was shown on the architecture slide. 14:51.000 --> 14:54.000 We integrate cool prop or interface cool prop. 14:54.000 --> 14:58.000 That's a fluid property database. 14:58.000 --> 15:00.000 We need that to calculate power. 15:00.000 --> 15:06.000 So we need to be flexible because a thermal sector is not very standardized in what we can get. 15:06.000 --> 15:08.000 And now time is up. 15:08.000 --> 15:10.000 Thank you very much.