Customer Showcase: City of Cape Town

00:00
Speaker 1
All right, City of Cape Town. The beautiful city of Cape Town. So the city of Cape Town, to give you a little bit of context, I’m not, certainly not the expert on anything water or city or infrastructure. To give you maybe just the background, the context. City of Cape Town, bulk water department produces approximately. I had to read this a couple of times, Kenny. A lot of zeros in there. Is it 1000 million? Which I realize is that a billion. Americans don’t call it a billion. So it’s a billion, I’m guessing million litres a day of water per day serving one of South Africa’s largest cities with a fairly geographically dispersed kind of infrastructure. So we have Kenny O’Kennedy and Rion Du Plessis, technical directors at Zutari and Rian associate at Zatari to join us on stage, please. Thank you very much.

00:50
Speaker 1
We can give them a round of applause. Sorry, first time on stage. Kidding.

01:00
Speaker 2
Yeah.

01:01
Speaker 1
Is it? Looking forward to it? Yeah.

01:07
Speaker 2
Rhian is also a technical director, I.

01:09
Speaker 1
Just want to say is real.

01:10
Speaker 2
Oh, he’s also a main guy.

01:13
Speaker 3
Chief cook and bottle maker.

01:16
Speaker 1
The subject matter experts.

01:18
Speaker 2
And yeah, it’s actually very recent that both of us are technical directors.

01:23
Speaker 1
Oh really? Is it?

01:24
Speaker 2
Yeah.

01:24
Speaker 1
Oh, okay. All right, I missed that. Sorry. Thank you. Thanks for that. Congratulations, by the way. Thank you for joining us today. We. We’ve got a couple of really cool visuals I suppose to go through when it comes to decision-making and empowering people with information. But I think what’s really nice about sharing some of what’s been done over really the last few years is kind of where it started. So maybe that’s where we should start is kind of how this came about.

01:55
Speaker 3
Yeah. Maybe I can kick us off. I think I’ve been involved with the decision support system the longest. Yeah. In 2017. Basically that was in the peak of the drought, 2017, 2018, the city had a need for having a centralized system, computer based, centralised system to really give them data driven information to support at that stage. A lot of people might have forgotten, but you know, we had quite a severe drought and were planning towards day zero and all that. And obviously for bulk water for the city, they need to manage the infrastructure. They need to consider, you know, what is the different scenarios, what will happen if the reservoirs do reach that level. Luckily we all know now that it never happened and yeah, and it rained again and that. But basically that kickstarted this need to have basically a data driven system.

02:57
Speaker 1
Okay.

02:58
Speaker 3
Yeah. And Zutari was fortunate Enough to be appointed to support them with this journey. So it’s been quite a long journey to develop a system that can give them insights into the entire water infrastructure. It is a like you mentioned, it’s complex distributed infrastructure with a lot.

03:20
Speaker 1
Of systems and silo data.

03:22
Speaker 3
Yeah and a lot of you know in the beginning stages there was a lot of data but a lot of, how can I say manual data. There was archive data. The city had a lot of years of historical data but it wasn’t integrated into a central system. The initial need was obviously more aligned to the hydraulic modelling and the network infrastructure but with it was you know, geolocation of all the valves distributed a couple hundred valves across the city on all the major lines.

03:58
Speaker 2
Dian that’s sitting there actually did a lot of that field work and collecting that data.

04:02
Speaker 1
Is that why he fled?

04:03
Speaker 2
That’s why they pass?

04:04
Speaker 3
Yeah, yeah it was like a couple hundred valves and all that but yeah so it was, so what was.

04:14
Speaker 1
The, what was the ask for Zatari were looking at essentially we want to get insights, build this dashboard and you guys, I mean you’re maybe a bit of quick background Zatari in terms of you know what you guys are specialists at and what you do.

04:28
Speaker 2
Well yeah so we Zutor is a consulting engineering company so we design and manage big construction projects basically so anything from roads to water to renewable energy. So what we do is actually we see ourselves more like a startup within, in Zutori but Zutori was helping at that point when we did this. When we started this Zutori was helping City of Cape Town of planning a trout. We’ve got a very good history of water and lots of water expertise. Not really but the, but our water team is quite well known and actually a lot of the dams and things were back, way back designed and built by our previous company. So the, so we started helping with that and then that’s also we realized no shocks, we don’t have any visibility of what’s going to happen if there’s no water in the reservoirs.

05:27
Speaker 1
So we wanted the visibility, the ability to get pre warning or get real time as near as real time levels as we can and be able to share that with people that pretty much anywhere or maybe a kind of central room. And that was the conceptualisation of the decision support system. The DSS.

05:46
Speaker 3
Yeah it’s basically getting a centralised computer based system that will give them insights into the assets and you know support them really in making that decisions to see what is going to happen across the network on depending on what the scenarios are.

06:03
Speaker 1
Okay.

06:03
Speaker 3
So obviously, our more the water hydraulic team supported on the, on analysing that data.

06:10
Speaker 1
Can we start with dessert? Can we, can we start with dessert?

06:13
Speaker 3
Yeah. This is the dessert. This is now where we, the current date. So we basically fast tracking seven years. So jumping to the dessert which is.

06:25
Speaker 2
Maybe, maybe we should jump to the next slide because I see the next slide. So that’s the overview of the, of the bulk of the bulk water system. So you can see it’s very dispersed. So that’s there’s dams, there’s reservoirs, there’s all the networks. There’s the valves at Diane Winton checked. It’s, it’s water treatment plants, its PRV pressure reducing valve each with its own.

06:49
Speaker 1
Kind of existing system infrastructure.

06:53
Speaker 2
Yes.

06:54
Speaker 1
Wanting to pull all of that together into one central data.

06:57
Speaker 2
Yeah, yeah. So just going back to the need obviously the need when you want to model your network and you want to see what’s happening with or what’s going to happen if there’s a day zero, you want to have as close as possible to real time data. So yeah, am I not holding my. And then, so that’s where the need came in for a skater and collecting that data in real time. So this is the skater part of that. And then on top of this there’s this what we call the DSS where we take that data. Now there’s another slide. Another slide, Another slide. That slide. Yes. So that is the modelling software alternate advisor.

07:42
Speaker 2
So it takes data from the historian, it pulls it into this modelling software and you can model if there’s some kind of event you want to model or you want to see how the flow goes through the city’s network or if there’s a contamination event then we would like to let Devil’s Peak Brewery know that they must not make any beer at that point.

08:07
Speaker 1
They’re not a large consumer, are they?

08:10
Speaker 3
But yeah, if we go back to the dessert slide, I think that.

08:13
Speaker 1
Yes.

08:14
Speaker 3
So obviously the main thing about the centralised system like this is you have lots of different data sources from manual capturing to Excel spreadsheets to SCADA systems. Obviously SCADA is one blocky there small aspect of it. Yeah, it’s one component of it. Yeah, but yeah like the structural assessments, there’s that dam safety reports. There’s lots of additional data that needs to be processed into the decision support system. And I Think that is where the using the app functionality really supported these applications in being able to pull in that data and then also do it in such a way where we initially rolled out the OT system, the SCADA portion of it, you know, we started with making tablets available more for the operators to go to industrial type tablets. Yeah.

09:05
Speaker 3
And then they can put in the, let’s say the valve open percentage etc and all that. And then from there the need sort of grown to being able to do these online data capturing instead of doing it in hard copy and then transfer it later. Yeah, but yeah, we’ve moved away now to even using WhatsApp.

09:28
Speaker 1
Moved away from using WhatsApp?

09:30
Speaker 3
No, no, we’re actually using WhatsApp. That is, that is actually Kenny’s. He likes it where, you know, where you can basically the operators can send a WhatsApp to find out various data or reservoir levels or whatever they want to know. They can utilize it as a tool to quickly get.

09:50
Speaker 1
Interact with the system.

09:51
Speaker 3
Yeah, interact with the system on, you know, they just quickly want to get this data or they want to see what’s going on here and really enables them to get that at short notice or quickly.

10:03
Speaker 2
Yeah, I think it’s, it’s about. You want to, you want to create this information model, data model and then you want to talk to the different stakeholders and you want to understand their needs and make it available to them in a way they want it. So some people, for them it’s just too much to go onto a website and go and check. They just want to sit at home and ask, okay, is my reservoirs full or what’s the damn levels at the moment? So that’s the idea. So I think the beauty of the whole UNS and building a data model that’s structured is that if you build it correctly then you can just add all these different things on top of it. So one of it’s what? It’s WhatsApp and Katle who is sitting there, he built most of the stuff.

10:50
Speaker 2
So if you want to ask technical questions, just ask him. Don’t ask me. And if you can just go back to the. Just the blocks. Sorry, our slide is not as nicely in order as the devil speak, guys. But so yeah, for me. So this is the front end of the data model.

11:13
Speaker 1
The primary interface. Sorry, the primary.

11:15
Speaker 2
Yeah, yeah. So the idea is that you have a data model and we busy. Some of it’s in a UNS structure, some of it’s not in a UNS structure. We’re Building this thing as we.

11:27
Speaker 1
So what are some of the components? Maybe when you say UNS. So in terms of your selection for the front end or the SCADA as we call it, the database, what are those technologies? What does that look like?

11:39
Speaker 2
The technologies. Okay, so maybe for me it’s better or easier to talk about the different data sources. So sure. So the data sources is obviously you’ve got all these sites and there’s a whole. There’s an old legacy telemetry system based on the Tetra network which. Which is obviously old technology and very slow. So that data is coming in. There’s. There’s data sources from MQTT brokers. We’ve got E1s on site that.

12:10
Speaker 1
And initially that was added fairly recently.

12:12
Speaker 2
Yeah, yeah, that’s actually now the last year and all if we’re working on that. So we moving this. We’re moving to this one central UNS types type system. We’ve got Paul Bride’s ZAN fleet reporting dashboards there.

12:26
Speaker 1
Hey.

12:30
Speaker 2
We actually said Grafana, but I think. I think that’s it for me. So this is ignition screen. The beauty of ignition is that it’s friends of everyone. It’s almost like Yaku. So it allows you to do bringing these different things. So that’s got me off track. There is unfree data. So that’s one of the biggest ways our children works that city of Cape Town has built recently massive projects. So all that data is coming through there or we’re using that as an interface then a big thing that Rian also talked about. I think actually that’s probably the biggest win that we’ve had is just moving away from writing stuff on a logbook or a notebook and then getting it captured.

13:20
Speaker 2
So the city and Alex is also here, they’ve done a great job in and we’ve also done a lot of training on it to getting the operators just to. To just capture it on a web form now. So it’s something as easy as that. But that makes a massive difference. So we have all that data as well in there and then obviously this is now the front end looking at that.

13:46
Speaker 1
Sorry, can you see you’ve got the kind of real time sources, if I can call it that, from different sites systems. You’ve got MQTT which was added subsequently a little bit later. Any other historical data that you’re bringing in from probably from some of the telemetry systems. Any funny kind of interfaces there or protocols?

14:07
Speaker 2
None of that I can think of. But the one big thing also is there’s obviously you’re doing your. With water. There’s lots of water quality stuff. So you need to do lab testing.

14:16
Speaker 1
Yeah.

14:17
Speaker 2
So the. Some of the on site lab testing that’s done on the web interface and then there’s also a limbs. The city’s got its own independent laboratory that. That data is also coming in here.

14:28
Speaker 1
Okay.

14:30
Speaker 2
And then I think. And that’s why we say we building this airplane while we’re flying it. So we realized as we. As we go in the city realize we’ve got a need for this. Need for this. And the beauty of having a flexible platform. Platform doesn’t. Doesn’t or allows you to just add on. You don’t pay extra is we realize there’s a need for. Every. Every year they do these structural assessments and dam safety reports. It’s a report someone goes and take a few photos. It’s very, it’s. You have to do it but it’s not a very difficult thing to do. So we use the ignitions app functionality. We built two apps for that. So now they can do that. They can do their assessment at. At the sites. We didn’t have the offline capability yet.

15:19
Speaker 2
We did vote for it on the roadmap, and I see they actually listened to us. So if you want to get. I think the roadmap.

15:25
Speaker 1
I think Travis saw your request on the forum and he said we have to do it for Kenny and Zatari.

15:30
Speaker 2
Yeah. So you can just speak to me if you want new features, any feature requests.

15:36
Speaker 1
Speak to you, speak to Kenny. All right.

15:38
Speaker 3
Yeah.

15:38
Speaker 2
So that’s the idea. You’ve got the. This is. This is our front end, but the back end we’ve got this sort of unified information model that we busy building and continuously upgrading so that we can hopefully the next time you invite us we’ll have more blocks there with all the things.

15:56
Speaker 1
Absolutely.

15:57
Speaker 2
And for me, each block is a user interface. So if Rian wants to see something in this format, just give him an app picker. Look into the data as you want to see it.

16:09
Speaker 1
Because not all data is valuable to everybody.

16:11
Speaker 2
Yeah, exactly.

16:12
Speaker 1
You know, data point to you is not necessarily valuable to me and you. The kind of the next step to figure out is how do you want that data served? Do you want it emailed every morning at 9:00?

16:22
Speaker 2
Yeah.

16:22
Speaker 1
Somebody else maybe don’t want that. They wanted the. Have the ability to immediately at any point in time see what it is. So we all have different needs from data at different times.

16:32
Speaker 3
Yeah. I think for Us the other thing just to tag on what Kenny said is the most important was having an open architecture with a platform that supports all these different data sources. Maybe didn’t mention SQL as well. So we have a few SQL data sources, a few web forms. The SCADA data MQTT is now recently added. But yeah there’s. And there’s different users. So it allows you. The guys that want to see their water modeling, they can look at that. The operators or the people that want to see what is going on the bulk water infrastructure terms of levels and flow and all that. This is basically for them to see the status of the network infrastructure.

17:17
Speaker 1
And the DSS I’m guessing is a physical location as well.

17:22
Speaker 2
No, it’s actually. Well there is a.

17:24
Speaker 1
From a control room point of view. Sorry there’s a couple of overheads.

17:27
Speaker 2
There’s not this NASA launching system. Not yet in some of the other department the city’s got that. But for the DSS I think it’s some of the. Some of. Some of those control rooms. I like it more that it’s a virtual control room. The idea is that anyone. Now this is also. We’ve got a domain and we’ve made sure that the people can access it from anywhere. So it’s a virtual control room in your pocket. So I can access this now because I’ve got the right credentials, I can access any data. So I like that more than the fancy control rooms. Although I also like that.

18:09
Speaker 1
Yeah, I agree. And from a. Interesting. From a WhatsApp point of view. Any specific reasons? So I see you making use of the Twilio module. I’m guessing the Twilio module for ignition or not?

18:24
Speaker 2
I’m not sure. You can answer that one.

18:31
Speaker 1
Okay, so it’s straight Twilio. Okay. All right. And any specific WhatsApp is fairly popular. I would imagine that most people kind of have WhatsApp. And you know why WhatsApp and not Telegram is it just everybody would have it kind of thing. Okay. All right.

18:47
Speaker 2
Yeah. This specifically request by the client.

18:50
Speaker 1
Yeah, yeah it is very popular interface. Fairly easy. I. I wanted to speak about this. So when you say you kind of do this simulation based on the historical data as well as the live real time data. So when it comes to kind of decision making capability based on this view, what are some of the kind of the decisions or actions that would be derived from something that’s red? That’s not supposed to be red. What does that look like?

19:20
Speaker 2
I was going to say you’re Going outside of me and Rian’s domain.

19:23
Speaker 1
Oh really? Okay.

19:24
Speaker 2
We are big. We are big team. I think that’s also the right when you as soon as you go from SCADA or just real time data to long term data you need domain expert.

19:35
Speaker 1
Yes.

19:35
Speaker 2
It’s one of the things you also realize so this is, this is actually a very like more domain expertise question you’re asking. But but I but we what’s nice about because we’ve got this data and the city is doing some really interesting projects in terms of reuse and there’s a big project that we’re involved with and they had this massive international panel that looked at okay can city of Cape Town actually reuse water? And they help us guide in all the making sure all the checks and balances are there. And then as part of that panel Alex and his colleague Lloyd actually presented what we’re doing with the DSS and they were so impressed with all the data that we have and that you can do all these simulations.

20:28
Speaker 2
So you can imagine that if you’ve got a reuse plant you actually need to know exactly how your network work. If there’s some contamination event you need to know what’s going to where. Where do you. Where must you which valve you must close and what you can what will work how will people be affected by that and that kind of thing. So I think just because and it wasn’t big we didn’t have this because of or so that we can do these more high tech water purification plants. It was just they already added now that allows them to do cooler stuff and everything.

21:08
Speaker 1
So and in terms of. I mean there was a lot of legacy systems and and devices and hardware and databases. We you had to bring in any kind of lessons learned things that you would do very different.

21:21
Speaker 2
There’s still a lot of legacy.

21:23
Speaker 1
Would think so any kind of approaches or designs that you would do different or I think for me.

21:31
Speaker 2
The and I just want to echo it Devil Speak guys also said is having a enlightened client.

21:39
Speaker 1
Yes.

21:39
Speaker 2
That allows you to experiment is very important because a lot of the other work that we do that are involved. We know you just go through. You’ve got these steps you go through and you design something and we build it where with this kind of stuff it’s fairly new and there’s no set recipe yet. Maybe in future there will be so if you and and you’re trying to. You’ve got these concepts of UNS and blah blah. And you want to apply to this, which is interest, which is a unique problem.

22:10
Speaker 1
Let’s build one. Yeah, let’s do.

22:12
Speaker 2
Yeah, yeah. So, and then you need to, we need to experiment. So we test this, we test this. So part of what we do is testing and we see it works. No, it doesn’t work. And then we go back and so and we’ve got a client that allows us to do that. I think that’s very important.

22:29
Speaker 1
Yeah, yeah.

22:30
Speaker 3
I think the lessons learned for me is that you know, you need to be able to change and the thing that we realized in this journey, so obviously there was a first a digitization journey before we can actually get to making any give providing any insights or cool dashboards or anything like that, is that you need a platform that allows open connectivity to any data source. You know, you can’t be locked in by specific platforms or protocols. You need to use open transparent protocols, open architecture and then build onto that.

23:11
Speaker 3
Because as we develop this further and as it’s growing, you know, we definitely know there’s going to be maybe additional needs or requests or you know, once the system is being used more and more, there will definitely be some additional use cases where there’s like, okay, now this portion of data has become available and because it’s available we want to add something new or we want to use it differently.

23:38
Speaker 1
I think you may very well find that bulk water system at the top may become City of Cape Town depending on kind of what other data and sources you can bring it.

23:49
Speaker 2
We hope, we’re hoping for that. So if there’s integrators that’s working on the especially wastewater side, then we can help you integrate with this.

23:57
Speaker 1
Okay. And long term data. Sorry, I had a lot of questions. The long term data store you currently use, I think you use Canary for some of the long term data.

24:04
Speaker 2
Yeah.

24:04
Speaker 1
So we’ve got the contextualised data.

24:06
Speaker 2
Yeah, we’ve got a historian, Canary historian. And then in the end we’ve got a postgres SQL database for the more aggregated data. Also the manual data.

24:17
Speaker 1
Okay.

24:19
Speaker 2
I think. Yeah. So just talking about the future. So the now we’ve, we at a place, we’re actually collecting a lot of nice data now we, when we, Alex also challenged us with that is we’ve got these, all these nice systems now I must actually show that the data is useful.

24:41
Speaker 1
Yes.

24:42
Speaker 2
So we’re Busy running a study at the moment, looking at, okay, we’ve got all this data, let’s look at, analyse that, let’s build business use cases of how we can use the data.

24:54
Speaker 1
Absolutely. Problem statements. And if you have the data accessible and available and contextualised, you can solve those problems. You don’t have to recreate the connections to try and find the right data. Lovely. Very nice. Any questions for Kenny Andrean? Comments? Questions, Andrew?

25:15
Speaker 3
Not allowed to ask questions.

25:19
Speaker 1
Oh, there’s a question over there. Liam.

25:25
Speaker 4
In terms of when you were developing out the UNS, how much effort did that take in terms of getting the data or into the same structure in a commonality with all these various sites that you’re integrating and the way they all individually are structured, how can you speak to the effort that’s involved in doing that?

25:47
Speaker 1
Lots.

25:48
Speaker 2
Yeah, we still, like I said, we’re still progressing with that. But just how to structure it is very difficult also because you have all these different sources. One of the things that we’re leaning towards is having a raw UNS. So you just get all your data and then looping it back into, and making it into a structured, more structured, nicely Iraqial side plant process area type thing.

26:21
Speaker 4
It’s like remapping it.

26:22
Speaker 2
Yeah.

26:23
Speaker 1
Yeah.

26:23
Speaker 4
Okay.

26:24
Speaker 3
Yeah. I think it’s. Some of it is also because of the history and the legacy of the system, some of these plants is I think cloth nec 80 years old. Yeah, 80 years old. So obviously some of the data we’re getting, there’s been at least 50 or 60 different tag naming standards over the years. So. Yeah, so getting it in the same structure is not that easy. But you know, there’s. Obviously, that’s part of the journey to get to that end state where everything is then nicely accessible.

26:58
Speaker 2
But the new plants that intake is working on, you guys must make sure.

27:03
Speaker 1
You guys must make sure.

27:04
Speaker 4
Give up my intern to do it. You just say, yeah, set up the screen and remap this.

27:14
Speaker 1
Good question. Very practical question. Cool. Anything else? Great. We are over time. Damn it. Damn it. We are overtime. Guys, thank you so much for sharing. I think that was valuable. That was really good. I’d like to personally see the kind of progression of this over the next couple of years. I think it could become something really, really valuable outside of just the bulk water supply. So good luck and thank you for sharing that. Give him a round of applause. Thank you.

27:41
Speaker 3
Thanks, Jaco.

27:42
Speaker 2
Thanks, Jaco.