VOICEOVER
Welcome to Up Close, the research, opinion and analysis podcast from the University of Melbourne, Australia.
SHANE HUNTINGTON
I'm Shane Huntington, thanks for joining us. Water, that precious resource and giver of life is scarce in many parts of the world, whilst being abundant in others. Yet in both cases the accurate and timely control and distribution of water is vital to communities, industry and agriculture. Compared to that of other technologies, research and development in the field of irrigation systems have remained remarkably under developed for decades, until now that is. On Up Close today, systems engineer Professor Iven Mareels joins us to explain how this crucial yet age old water transport method is intersecting with modern information technologies to boost irrigation efficiencies and reduce waste. Professor Mareels is Dean of the Melbourne School of Engineering here at the University of Melbourne, Australia. Welcome to Up Close Iven.
IVEN MAREELS
Thank you Shane, it's a pleasure to be here.
SHANE HUNTINGTON
Today we're going to be talking about some of the implications of the use of systems engineering. But before we do that, I thought we should start with a basic description of what systems engineering actually is.
IVEN MAREELS
System engineering is indeed a very much abused term and it's used for many different things. But from the point of view that I'm taking, systems engineering is basically a field of study that looks at the interaction between systems. Systems in that sense is the very broad word for anything that takes information from the outside world, transforms that information and then takes an action.
SHANE HUNTINGTON
How old is this area of systems engineering, and where did it initiate?
IVEN MAREELS
It originally started early 20th Century, and probably had a different name at that time. But Norbert Wiener was a famous professor of Mathematics in MIT, who gave the name cybernetics to it. But since then we have looked at the word systems engineering, probably early 50s. So you could say probably about 60 years old.
SHANE HUNTINGTON
Can you give us an example of some of the problems that a systems engineer would typically be tackling?
IVEN MAREELS
Well one famous problem that probably everybody will remember was the landing of the man on the moon. The whole computation of actually bringing the module out of the orbit, putting it on the moon, was actually done by systems engineering.
SHANE HUNTINGTON
Iven, can you give us a bit of an idea of how the methodology behind systems engineering differs from that of other fields.
IVEN MAREELS
Systems engineering has basically developed along two lines of thinking. One is the classical bottom up approach which I would say is the scientific method. You try to find a smallest entity of understanding in your overall object and then you build the object up from all these little bits and pieces. That works very often well, if what you're looking at is not too complex. But what we find is that when you get to very complex objects, that that bottom up approach often fails because it's just getting too much clutter and too much detail. So coming from the outside and seeing what actually is the objective of that object, what does it actually want to do, how does it actually work, understanding from an outside point of view if you like, what a system is doing, that is a top down approach. Systems engineering has a good mix of understanding, when to do top down and when to do bottom up and how to interact between the two basically.
SHANE HUNTINGTON
Much of your work has been focussed around the application of systems engineering to agricultural irrigation. Before we get into that, can you give us a little idea of how this sort of irrigation is done traditionally? What the mechanisms are and I guess what some of the challenges are.
IVEN MAREELS
Sure. Irrigation basically is distributing water in time and space, differently from the way nature provides it. It has been practised since 3000 or 4000 years BC, since we first found the first settlement if you like. Typically you first harvest water in a large basin, a dam of some sort. Hopefully at a reasonably high location so that the gravity that's then available to it will be able to be used to drive the water to flow to any different location you like. So from the dam you build a set of channels and that set of channels will go and take the water from the dam to, say , the farm and in order to make good use of the water along the channel you will put certain structures in place that will allow you to regulate the flow. For example, divert it to another channel or stop the flow in cases that you don't need the flow anymore at a particular location. It has been done since the Sumerians in Babylon 4000 BC easy. The Romans perfected it to some extent, at least the civil engineering part of it. They understood most likely how to build the complete system. Really we only have improved the civil engineering part of that. Even today when you go and look at large irrigation districts, even the ones that are being constructed right now, are still constructed very much along the same principle. Use gravity to disperse the water, build your channels, put in structures that can change the flow and regulate those so that you can deliver water to a farm. It's still done today, very basic, very simple civil engineering really.
SHANE HUNTINGTON
What's the major drawback with this approach?
IVEN MAREELS
Well as long as there is water a plenty, there is no real drawback. It does what it has to do, it's very simple it delivers the water to the farmer. The drawback comes at a point where you have lots of competing interest for the same amount of water, and we actually start having shortages. Or where you start seeing that farmers are using too much water and then you have drainage problems. Well when you start hitting the boundaries, I would say, of how the system can work then it really is in trouble. Because there is no smartness about it. Once water is dispatched, there is no way of calling it back. There is no way of knowing where it actually is going until it's taken out by a farmer somewhere, or it goes somewhere, the end of the channel maybe in a desert or maybe in a pool. But there's very little knowledge about where is the water, what is it doing, who is using it. All of that information that you would like to know when you want to manage a resource is not here.
SHANE HUNTINGTON
So presumably then when you're looking at the competing priorities of individual farmers for example versus the system as a whole, these traditional mechanisms are essentially unable to balance those competing priorities. Is that right?
IVEN MAREELS
That's absolutely right, spot on. If you have competing priorities, if you have even competing farmers, it's very difficult to be able to serve both. Because there is always one that is upstream from you, and therefore has a first call on the water if you like. If you can't manage that then you can't manage the whole system.
SHANE HUNTINGTON
Do we have an idea in some of the larger systems here in Australia for example just how much water is lost as a result of this poor management scenario that we find in these traditional systems?
IVEN MAREELS
Well that in itself is a very interesting question because in effect we don't measure it very accurately. So even when we say efficiency is around 70 per cent, which is a typically quoted number for Australia, we actually find it very hard to justify that or quantify that perfectly. Other places, like in China will quote quite openly that they are probably only 30 per cent efficient. Most of the world the number is somewhere between 50 per cent and 70 per cent.
SHANE HUNTINGTON
This is Up Close, coming to you from the University of Melbourne, Australia. I'm Shane Huntington and our guest today is professor Iven Mareels. We are talking about systems engineering and how it applies to recent developments in irrigation technologies. Iven, give us an idea of how you went about applying systems engineering to these problems in irrigation.
IVEN MAREELS
Well first it's actually a strange thing to have a systems engineer looking at irrigation. So what actually happened was that a company Rubicon Systems Australia, came to ask me a question and said, look, in Northern Queensland we have these channels. They are being automated and we can't really make the automation work. The automation came from the United States. The people that were installing it are going bankrupt. The Queensland government wants to do something better with it and now we have been given the task to do it. It's hard work, it doesn't work. We get waves on the channels, the channels don't deliver the water on time, well they had troubles basically. So they asked me, can I have a look at it. I found it quite an interesting question in a way because it was definitely not a simple system. There were lots of channels involved, lots of regulators involved it was a fairly nice irrigation district that they were talking about. We said, well yes, we'll have a look at it and see what we can do with some of our techniques. I told the company that we would look at it from a systems engineering point of view. That we would probably approach it from a very different point of view than an irrigation a normal hydrologist would approach the problem. We were first going to do a top-down analysis on the whole system. Which they found themselves interesting, because actually nobody had ever done that for them. So we went through what are the specifications what actually has to be achieved with such a channel system. Things that we would like and things that we wouldn't like. It all came down to basically getting the right information at the right time. So that was the first place we started looking at, is how do we get correct measurements of water flow/water heights in the channels. Once we got that, I had a feeling that from that data, from the system engineering point of view, I would be able to deduce the type of models, system models, things that we put in the computer that we can compute with, that will allow us to predict the water flow in the channel. Then from the predictions of the water flow, I believed that we would be able to start making management decisions. So that was the starting point. We started in 1998 with that question basically. The rest is basically history.
SHANE HUNTINGTON
Physically when you're down on the farm or looking at these channels, what does this look like now? I mean I think some of our listeners will have this image of this open channel irrigation system. But obviously to add this level of control you've changed a lot.
IVEN MAREELS
Yes, so we didn't actually change a lot to the channels themselves. So when you look at them typically irrigation channels will still be a ditch in the ground essentially. Whether it's concrete lined or plastic lined or just soil, it will still be a ditch and they vary in size. Some of them are 30 metres wide and three metres deep and go for kilometres. Others are only 30 cm wide and a small ditch. But that's not the real issue for us. For us it's more to know where is the water. So basically the structures that people are using to stop water flow or to increase water flow we changed completely.
We made them quite modern we make them such that they don't leak anymore so we can now precisely measure the water flow through the structure. We know exactly how much water is in front of the structure and behind the structure in terms of depth. These structures are completely redesigned to fairly high precision engineering and allow us to have quite good precision about both water levels and water flow. That’s the starting point.
So when you look at the new irrigation which have these technology, they look very different from the point of view that you have over everywhere these gates that we call them, these structures are completely changed. They look very modern they have a power source now, a solar cell associated with them. They have some communication channels associated with them, so you can see it's very different from that point of view and it is indeed very different it's very modern and it's quite an instrument actually.
SHANE HUNTINGTON
We spoke earlier about that interplay between the priorities of the farmer and the system as a whole. How do these new gates and the systems that you have implemented allow for that prioritisation to be managed?
IVEN MAREELS
Okay so from the gates we get basic information about the quantity of water, both the flow as well as the storage in the channels, which is significant. So if you start doing it across a complete channel you have as a manager at hand, completely water balance in time and in place, completely distributed at your fingertips. When a farmer comes along and says, I want water, I can then compute if the system is able to deliver that water and at what time it can deliver that water. Because it has enough understanding of the physical properties of the system to be able to decide, yes, this farmer can have the water in, say, two hours' time. Once that computation is done and it goes very fast, the system, basically will alert the manager and say look, the farmer is asking for water. Physically it's possible to be delivered. Then a few other things can go on depending on how the manager wants to manage the water. It depends on the priorities of the farmer, how much they are wanting to pay for the water, what type of priority they want to attach to the water. Taking it all together, we can then decide what is the priority scheduling for that farmer and then implement it. So without any intervention of human beings, the system or the channels will start working together to deliver the water as ordered by the farmer at the time they want it.
SHANE HUNTINGTON
Presumably this difference from the old system where there are quite significant delays in the delivery. I can imagine that would potentially almost cause the death or over-flooding of crops and quite significant financial loss.
IVEN MAREELS
Correct. So in the old system what happened was the farmers put in their requests for that week and typically they are for the week because the ordering delays are typically between four and seven days. They have been computed based on where the farmers are with respect to the major dam. Then that person will then start ordering water from, his superiors and say, look I need so much water and I'm going to distribute it in such a way. But it's all done manually based on experience. It's much less flexibility. Farmers not only have a very long lead time, a minimum of four days. They also have to order water in very large quantities like 24 hours, 48 hours which makes it very difficult to run a farm basically or to be responsive to the climate. The only exception they make is if it starts raining, the farmer is allowed to stop the water that has been ordered to come onto his farm, which is painful as well for the person that has to manage it because that means that all the ordered water will just disappear and can't be used anymore because the system can't respond to an event like that. So they are a few of the difficulties that exist in the present system that you can avoid when you have an automated system.
SHANE HUNTINGTON
Iven, I suspect many people when they think about this, they think about low water conditions, drought conditions and how to distribute this precious resource. But in Australia in particular, we also see the opposite. We see substantial flooding in parts of the country. How does this new systems management approach allow us to better deal with flood conditions?
IVEN MAREELS
That's a good question actually. The flood conditions, just like low flow conditions are basically part of what the system can handle. If you had a manually operated system and we're talking here about channels that are very long. I mean some of our channel systems have links of 8000 kilometres of channel in theory. You have to go one by one through structures to try and change their flow in order to cope with the flood. The flood just overruns you and you can't cope with it. On the other side, if you know a flood is going to come, because you know that your dams are getting overflowed and your river levels are too high, you can do a pre-emptive strike with an automated system. You can take water out of the dam, disperse it as quickly as you can through the system and you can open it to maximum flow bypass most of the farms and basically dump it out of sight. That way you can avoid a flood much more efficiently. In actual fact, during the first part of the trial we had, there was a condition like that that happened. It was a future rain event, flash flooding occurred, our whole system shut down, avoided all farms to be flooded. What was more important to the farmers anyway, not only the farmers that did not have the automated system had wet feet and our farmers did not. After the dry period started up again the system could react again immediately and put the water onto the farms where it had to go. So the system that was manually operated took three days to a week almost to come back to normal. Whereas the automated system just started up again as [if] nothing happened. That was a huge difference. Basically that was the thing that convinced the farmers that that's what they wanted.
SHANE HUNTINGTON
I'm Shane Huntington and my guest today is Professor Iven Mareels. We're talking about systems engineering and new generation irrigation systems here on Up Close, coming to you from the University of Melbourne, Australia. Iven about what portion of Australia's fresh water is actually used in rural settings?
IVEN MAREELS
Australia is no different than the world here Shane. Roughly 70 per cent of all the water that is used in Australia goes to food. I often say food is packaged water, 70 per cent. 30 per cent is for the rest and that means about 12 per cent for the cities and about 18 per cent for industrial use. That's roughly what it is in all the world except for some countries where probably the economy is more industrialised the percentages go down. But you will find rarely a country that is less than 50 per cent rural.
SHANE HUNTINGTON
About what portion of Australia's irrigation in these rural settings is actually now covered by this new systems approach?
IVEN MAREELS
The largest district is the Northern Victoria District called Murray Water. So that's a channel which is about 7000 kilometres from top to bottom and has about 20,000 farm outlets on it. It's now been completely automated. There is still a way to go. We have a few other smaller systems being installed as well. The Coleambally Region, the rice region of New South Wales is also fully automated. We are looking at some instalments in Queensland as well.
SHANE HUNTINGTON
So between Queensland, New South Wales and Victoria you essentially have all the states down the Eastern Seaboard of Australia.
IVEN MAREELS
Correct yes, that's basically done.
SHANE HUNTINGTON
Now Iven, what about the rest of the world? I mean water usage is obviously a major problem internationally, both in scenarios where there's excessive drought and rain. Are these techniques being applied elsewhere at this time?
IVEN MAREELS
Yes so thanks to a number of initiatives that the Victorian Government here in Australia has helped a company to set up internationally. We have now our technology being applied in the United States and there is a Rubicon United States for the moment that also services South America. and And there is a large irrigation district, actually larger than the Victorian irrigation district that we are automating here, that has shown a very big interest and we've done the first two pilot projects over there. That's Imperial district in Southern California and there is also a big interest in Chile with respect to the technology. Also the new developments in Brazil just north of the Amazon, the people there have expressed interest too in acquiring technology. There is a tendering process going on in India for one of the largest irrigation districts in the world to automate the full main channel and also on farm distribution. Again our technology is being considered as being in the lead for that application. In China there is a huge interest because probably the biggest shortage of water will happen first in China. The government there have basically said that they want to implement similar technology across most of the northern irrigation plains in China.
SHANE HUNTINGTON
We've been focussing so far very much on rural systems. But water distribution in large cities like Melbourne here in Australia is also something that is of deep concern. I guess those systems of varying age give you varying conditions. How does a city compare to these rural irrigation systems?
IVEN MAREELS
So from a system engineering point of view it was logical for us to start with rural simply because it's so much larger. But you're absolutely right, cities still account for about 10 per cent of the water and it is rising as the population becomes more and more urban dwelling. The issue there really for us is one of age, as you point out the channels don't exist of course, they're all closed pipes in a network. The speed with which we deliver water in an urban system is much faster than what can happen in rural. Urban channels are slow, it's only potential energy available, but here in the city we of course build up the pressure. We have water towers and other means of putting up pressure higher, which allows us to deliver water much more quickly. We don't really worry about storage in the system itself. The pipes themselves don't store any water. So it's a very different dynamic, it's a very different type of environment to operate in. But because of the difficulties that people have encountered over the time, it's actually becoming interesting to look at it as well from a systems engineering point of view. The system's age, like you point out, ageing means that the main pipes probably are leaking very badly and the only way we know about that is when suddenly a geyser erupts in the street. Then we know, yes, well somewhere there's a leak and we go digging for it and finding it and repairing the pipe. Very costly to everybody involved, not in the least the company itself. If you could predict where leaks were, and identify where they were then we could do pre-emptive maintenance. We could avoid such situations and we would probably have a much better understanding of where our water is. Because at the moment even in cities like Melbourne 10 to 20 per cent of the water is probably completely unaccounted for. If you go to older cities like London or New York, you're probably looking at percentages closer to 30 or 40 per cent of water that is unaccounted for. If you go even to modern cities like Malaysia or you go to Indonesia and Jakarta, it's not unreasonable to expect 50 per cent of the water in those cities to be unaccounted for. It's partly because again the same thing, we don't actually measure water all that accurately. In our society we have a meter at your home and you can pay for your water there. But that doesn't tell us much about how the water got there, it's only the water that got into your house. Even there, your water meter can be turning over day and night because there's a leak in your house and you wouldn't know for most of the time. Because it's not expensive enough so you wouldn't take notice of your bill. But also for us it's actually important to know where those leaks are. We don't have enough measurements on the pipes in the networks to be able to do that. So to actually start looking at the whole urban network, we have to start again from where we started with the agricultural system. There we started with measurements here we have to start with measurements as well. You have to have a density of measurements which is far higher than what people have implemented so far. That's a hard sell for the moment because, well for most of these systems, they work reasonably well. It's not like we don't deliver a good service. It's just that again the same thing, in my opinion, we don't deliver a service with the same type of accountability that I believe will be necessary in the future.
SHANE HUNTINGTON
Presumably to make those sorts of measurements and to make those determinations, you have to be able to determine the pressure of water at a range of positions in the system. Is that right?
IVEN MAREELS
Yes that's absolutely right. So again putting in new sensors, pressure sensors is one thing, flow sensors is another. It's difficult to do it in a piping network, we don't like to interrupt pipes and put flow meters on top of it. Pressure sensor is probably easier, but nevertheless a pipe network is normally underground and therefore it means digging holes and putting a sensor in place is a very expensive exercise. Not because of the sensor itself, but simply because of the earth works involved of people costs involved, the interruption time involved. So when you suggest to an operator that they suddenly have to install a few thousand sensors in order to manage their network better, they start counting the cost very quickly and the revenue for the moment isn't there. So the economic argument is a harder one to make in the urban area.
SHANE HUNTINGTON
You have in the farming setting the ability to put in these gates quite readily. What would these gates look like in an underground pressurised water system in pipes for a city like Melbourne?
IVEN MAREELS
So in principle they become like water taps but the taps will be a bit bigger, they will be sitting in line with the pipes. Some of them are already there, but they will have to be automated. So we will have to know exactly how much they are open, what pressure drop is across them, how much flow they let through. That's the type of thing that we have to start knowing about all of these environments. Surprisingly maybe, what will happen first in urban environment is sewerage network. Because sewerage networking, especially because of our making more efficient toilets and everything else, actually are over designed. They were designed for flows which are much larger, probably about four or five times larger than they actually are at present. So that allows you to create a higher density in the housing, but that then creates some other problems for these networks, how to cope with the sewerage essentially and to manage sewerage. That's probably one place where there's an economic argument, because it's a higher density of occupation, that can lead to automation of sewerage networks to protect them, to make sure that they don't interact with the installed network and that they actually deliver the quality of extraction of water that you want.
SHANE HUNTINGTON
Iven, this requirement to move towards these systems in many countries I can imagine there's a bit of a futurist's argument in that we can clearly see the benefits in the future, but driving governments and private operators to expend on such systems in the here and now is going to be a hard sell. How do you intend to go about that given that the need in the future for us to have done something at this point will be so great?
IVEN MAREELS
You're absolutely right Shane, and actually talking to you right now might actually be a step in that direction. I believe there is a need for doing these things to be much more accountable for our resources and therefore essentially you have to have a policy change. Policy change means that politicians have to care and that people have to care about it as well. Otherwise politicians won't care. So if people start understanding that water is actually a very precious resource, and that there is actually enough water if we manage it well. But we have to manage it well and we are starting to see the boundaries of what can be done with planet earth and how much water we can actually harvest and use and so on. We're supplementing it with expensive water, like desal water. All of these things are part of the system and I think recycling has to be put on the table as well as an option for us and how we recycle and how do we manage equality and all of that has consequences. Because face it, water is the most important thing for life. Having clean water in the home is probably the most important invention that engineers ever have made for humanity. It's probably the one that gave us the most increase in longevity ever. If you don't maintain that basic quality, we actually undermine the basic quality of life. Maybe it sounds dramatic but I really think we have to pay attention to this. The only way we can do that is by starting to manage it much more precisely than we have done before and for that, I will advocate yes, let's put in the sensor networks, let's put in the information gathering points that we need in order to put in place of everybody a clear view of what is it that we do with our water? Do we do the best? Are we actually using and managing our resources as efficiently as we possibly can? We never can answer these questions and never find a better way forward unless we start metering it.
SHANE HUNTINGTON
Iven there's obviously a big contrast between the way in which this engineering is done and the engineering of the past. As an educator at the university how are you going about instilling this sort of more holistic view into your engineering students to make sure that they do produce systems that cater for both the needs of the consumer and the environment and the whole system.
IVEN MAREELS
Very true. In the past engineering has developed very much in a number of silos, electrical, mechanical, civil engineering to name but a few. Really when you talk about water, it's a system that requires civil engineering thought, it requires mechanical thought because all the pumping and everything else that's involved. It requires information technology and software engineering to manage it properly. They all have to be brought together. So we can't teach mechanical engineering in isolation from electrical and from software engineering. One of my dreams is to really come back to the renaissance engineer. The engineer that has a good understanding of what are the basic principles that actually establish engineering. I don't see engineering in any very specific limited sense. I see engineering as being creative and you are creative with the physical or the limitations that are put in place in front of us by nature. We exploit them to do something for humanity. That's what I would like our engineers to be like. We're starting to put in place that type of education where we bring all of engineering together, let the students choose how they want to learn within that environment and then maybe develop later on a speciality. But they have all got the same basic foundations to start working on. Then these problems don't become such problems anymore.
SHANE HUNTINGTON
Professor Iven Mareels, Dean of the Melbourne School of Engineering, here at the University of Melbourne. Thank you very much for being our guest in up close and giving us such a great understanding of how systems engineering can be applied to irrigation.
IVEN MAREELS
You're welcome Shane, thank you.
SHANE HUNTINGTON
Relevant links, a full transcript and more info on this episode can be found at our website at upclose.unimelb.edu.au. Up Close is a production of the University of Melbourne, Australia. This episode was recorded on Thursday 31 March, 2011. Our producers for this episode were Kelvin Param and Eric van Bemmel. Audio engineering by Gavin Nebauer. Background research by Dyani Lewis. Up Close is created by Eric van Bemmel and Kelvin Param. I'm Shane Huntington, until next time, goodbye.
VOICEOVER
You've been listening to Up Close. For more information visit upclose.unimelb.edu.au. Copyright 2011, the University of Melbourne.
© The University of Melbourne, 2011. All Rights Reserved.