Smart Water measurement


(Flavio Maeda) #1

As the result of the Meetups we did in Sao Paulo, we were approached here by the founder of a start-up who brought an interesting use case to discuss with us.

This startup develops an App to be used by the people who lives in the residential buildings in Sao Paulo (Sao Paulo is a vertical city so the majority of its population lives in buildings). This App has several features such as to promote the communication among the neighbors and the building manager.

He saw lots of other IOT applications for this App. Sensor nodes spread in the buildings could give lots of real time information to the inhabitants of the building, such as the level of the water reservoir, the temperature of the swimming pool, security alarms, etc.

One of the most impacting sensor however would be the water metering to measure the water consumption of each apartment. Sao Paulo is facing a water crisis and the entire city is being challenged to reduce the water consumption. However, most of the residential buildings (the old ones) does not have individual water measurement (one meter per apartment) but only one central meter that is used to measure the total water consumption of the entire building.

This startup will provide this application to more than 300 buildings in Sao Paulo and they are willing to fund the installation of gateways in each of this buildings to provide IOT connectivity not only for the inhabitants of the building but for all the community in the surroundings of each gateway.

I saw in the LoRa catalog (that @wienke sent to me) where there are many LoRa sensors listed and specially designed for water measurement but in order this use case be financially viable it would be better that the water measurement be done without the need to install a water meter in each apartment (breaking the walls to install the meters), but some type of sensor that can read the flow in any other way.

If anyone knows any type of LoRA ready sensor that can do that please let me know!


Simple fluid flow indicator
(Nestor Ayuso) #2

I work at mirakonta, we are listed in the LoRa catalog. Our product is a clip-on reader attached to the water meter.

There are solutions to read water flow directly in the pipe tube without a water meter. They are "ultrasonic flow meters" are very expensive, only used in municipality water supply networks or industry, not in residential use.

Try this kickstarter project: https://www.kickstarter.com/projects/825947844/fluid-the-learning-water-meter


(Flavio Maeda) #3

Thanks @nestorayuso.

I saw your clip-on reader in the catalog. I believe it will be very useful in Brazil since 99% of the water meters installed here are the old ones and I understand your device can be plugged on it and be used to read the measurements. What is the price of this product?

Regarding the one in Kickstarter (Fluid) do you believe it can be adapted to use LoRa rather than WiFi? However, I believe it is too expensive for the Brazilian reality.

Thanks!


(Patrick) #4

See these too — http://www.libelium.com/smart-water-ions-sensors-calcium-fluoride-chloride-nitrate-iodide-lead-bromide-cupric-silver

It is built on the waspmote — and can be developed independently of its enclosure/setup.

Cheers,
Patrick


(Brendan O'Gorman) #5

@flavio @nestorayuso Following this, from south Asia. The concern we addressed at our meetup was in regards to Dengue Fever that is spread by mosquitoes. They have teams that go about looking for standing water and even issuing fines to home owners who have left standing water on the lots which are breeding mosquitoes.

I can envision these inspections being augmented or assisted in setting up a sort of surveillance to monitor standing water in certain places that are problematic and informing home owners automatically or dispatching spraying at the most appropriate times/places.

The other issue is early detection of flash floods, when flash floods hit they often knock out power and communications making the warning even more difficult to notify. Flash floods and mosquitoes - Brazil faces similar challenges?

How could this relate to the sensors and applications being discussed here?


(Lawrence Griffiths) #6

@Brendan what you might want to look at is Soil Moisture sensors.
The professional end of these are used in precision agriculture about $100 per sensor.

At the maker end of the market price range - indoor use
http://www.seeedstudio.com/depot/Grove-Moisture-Sensor-p-955.html
To - for out-door $40 a good price this uses a good quality digital sensor
http://www.seeedstudio.com/depot/Soil-Moisture-Temperature-Sensor-p-1356.html

One question what area are we trying to cover at each location?
You can create a wired bus of sensors uses SDI-12 but these are not cheap.

Lawrence


(Flavio Maeda) #7

Hi @brendan. Flash Floods and mosquitoes are a big problem in Brazil too. Dengue Fever is a huge problem indeed and with the summer season coming many people are trying to find ways to apply technology and innovation to mitigate and prevent the damages. In Sao Paulo city I know an innovation lab of the City Hall of Sao Paulo which is planing to use Drones to detect opened water tanks (located on top of the houses) which are the main sources of proliferation of the mosquitoes. They want to detect these spots using drones equipped with cameras and using image recognition software to alert the authorities and the house owners.

Any idea is very welcome. I was still unable to think about an IoT application that can help on preventing the fever, standing water, or the mosquitoes, that can be affordable to be widely implemented in the city since the possible areas of accumulation of water are small and spread over the city. Putting sensors in all these places would not be doable. What do you think?


(Brendan O'Gorman) #8

@flavio I have a few ideas I will share maybe they spawn ideas with your team.

For water tanks - Using ultrasound or infrared detectors submerged, pinging air from the water - to monitor levels (+ add temp sensor and chlorine sensor) encased on epoxy resin or whatever.. ideally we would just drop these in the bottom of water tanks and they start plotting water levels and determining normal usage patterns.. we take these and match them against the known conditions that are conducive to breeding mosquitoes. We give a green, yellow, orange, red sort of warning system. The problem with both drones or even sensors is scale.. more of a problem for drones. How many roof top water tanks in Brazil.. and this does both.. mosquitoes and the water consumption app you mentioned in buildings. Drones only do mosquitoes and other surveillance.. drones could do checking standing water around people houses perhaps it could work best as a combined approach.

But I feel dedicated fixed sensors are more practical for more static environments, such as out in the hills or inside the water tank, or actually even around the house.. once the detections and sensors where refined enough, not that far away I suspect.

It's also important to consider the acceleration rate of technology now.. does it favor the drone investment or the simple sensor device investment?

Floods and and mudslides would be also rather simple detections?.. could use sweeping laser pulse and optics package - water levels, changes in metrics .. knowing the conditions that lead to the flood or mudslide..monitoring the conditions upstream at hills and mountains and all the way along. Maybe one device every 15 Kms along river beds for flood and multiple sensors and (perhaps interconnected with ground moisture sensors as @iikaw mentioned (thnks.)) at areas at risk for mudslides?

Like the mosquitoes we actually know conditionally, where they will be and when they will be a risk, but we lack the fidelity in monitoring and early warning. Especially with floods, avalanches and mudslides, because as they happen they destroy the power and communication systems and ability to warn/notify, we can overcome that also.

Not refined , but it's something like that I suspect. I don't think it's a drones, how many $5-10 sensors could you buy with one drone? How many drones would you need to monitor 50 million water tanks effectively? Each sensor will monitor one tank perfectly all day everyday for years.. is there a cheap chlorine sensor? I think it's chlorine people are supposed to use? We can send notifications to home owners to self service.. offset costs of manual inspections..etc..

We could use other sensors or even optics to keep an "eye" on the ground within a limited range in indicative areas which we could identify puddles forming (and drain blockages) and standing with temp info and weather info, calculate probability.. Eventually this system could be interconnected with other environmental sensors and weather services to give long range accurate predictions (and better prediction modelling). I don't think these electronics are expensive, just numerous.

I also have had the idea for a while about easily being able to detect and plot pot holes across the road networks.. don't even need to make a device, just a small app that runs in the background of smartphones.. fly me to Brazil :smile:

In all situations we are monitoring the physical environment for physical changes / patterns.. we have all kinds of low power low cost sensor electronics to achieve this - now we have the Network to cover it. Though it sounds like a statement.. this is actually just a big long question - what do you think?


(Flavio Maeda) #9

@Brendan

Thanks for all the ideas. Let me explain a little bit more the problem with the water tanks. The picture below shows the type of common and low cost plastic water tank. There are millions of them on the roof top of the houses, specially in the poor areas.

The main problem is when the plastic cover of these tanks is removed (either intentionally or by a wind storm) and the mosquitoes have access to the standing water inside the tanks.

See this other picture of a poor area.

What the drones will do is to detect the uncover tanks to notify their owners.

Putting sensors in all the millions of tanks is impossible specially in the poor areas.

Floods and mudslides would be easier since there are high risk areas where monitoring and early notification is very important.

It will be a pleasure to have you here in Brazil. If we have a killer idea we can easily find a sponsor! :smile:


(Brendan O'Gorman) #10

@flavio Cool.

Why not just go straight to Fortlev with a simple ~5$ lid detection device?.. not only detect if open, but locate (if it blows away) have them Give it away as promotion for their part in the fight against dengue (PR) .. or the challenger companies in the market.. they would be even more eager to make that announcement. Their new "Smart Water Things". So that could set the trend for a higher product standard, solving cost and logistics.. even if it adds a few $ to the cost.

I have another idea, will think on it some more before saying more...

Update; on thinking of the lids again.. this device could be super simple/cheap actually conceivably.. it could be powered only on exposure to light.. and all it does is transmit a intermittent signal when energized (exposed to out side light through a single tiny photocell) It's essentially a tiny solar panel?

Another variation of a distributed device network for this could include aspects of delivery/placement of devices by drone.. since it is intended drones will be out and about inspecting.. this gives a development fork towards the drone systems development, such as device payload carrying / delivering, by knowing where all the sensors are or are not and what their state it is, (the 'drone') would be able to derive the best search patterns making them more effective, efficient. It could figure out by 'itself' where it should check next basically. ? Or it could tag tanks with open lids and 'see' when the lid has been re-affixed, from the device monitor. (just collab'ng :wink:

Miniaturization and cheap cheap sensors could open a world on new ideas things we say today are impossible or unpractical may become feasible soon.. adhesive sensors the size of small NFC tags, could be down to less than $2 and hundreds of them could be carried by drones in rolls and placed like price stickers are placed on products in shops.

You don't think passive monitoring of the accelerometers in smart phones to identify and plot changing road conditions is a killer app (when holes and tires collide they produce a distinct motion and sound signature)? Your team could probably build it tonight and announce it in the morning lol.

Update: the other idea' ..object recognition devices make sense for this, basically it is why they would be flying the drones to identify the specific types of objects (containers with or without lids) there are many patents and such flying around , companies specializing in facial and vehicles or specific types of pattern detections and different ways of doing it. These can be placed around in key vantage points of roof tops effectively covering a city seeing most of the rooftop storage units. How many devices are required depends on the optics, you don't need particularly high resolution in daylight but need long focal depth, auto-focus, auto-zoom .. the better the optics the less units required. Perhaps there is someone in your team or another who is familiar with this technology who could help determine feasibility .. I have working familiarity with AdMobilize.com and spent a few months applying there tech to mobile marketing engagement.. it works.. and it's pretty exciting what can be done with it. All processing is done on-board.. so no images are transmitted only encrypted data, can be OTA updated with new detection patterns (birds, bats, Santa).

Where this tech is challenged is fast processing, such as providing analytics on a group of people at once quickly to determine individually their age/sex/ethnicity/height/etc.. as they walk by.. but for someone looking at a signboard.. its very fast and clever at identifying these characteristics. SO the roof top challenge is not faced with that condition, in fact they (detection devices) can take their time to figure out exactly where each container is and get better at spotting them, by color, and dimension or even partial brand logos, water surfaces, etc.. this network could play nice with drones. Some related links newer and older.

http://www.freepatentsonline.com/y2014/0003671.html

see also Wikipedia - Glossary_of_machine_vision

http://www.neurotechnology.com/sentisight.html


About the Malaysia category
(Flavio Maeda) #11

@Brendan I am sorry for the late reply. Your ideas were so good that it is taking time for me to digest everything. I liked the idea of a small and low cost light sensor to be placed by flying drones inside the tanks. My wife also gave me the idea of the same sensor detects movement of the mosquito's small worms. There is a project going on right now for visual recognition of standing water spots in Sao Paulo (such as opened water tanks). I will talk with them and brainstorm around these IOT ideas. Thanks!


(Brendan O'Gorman) #12

@flavio Cool. Actually my wife also suggested the same, about detecting ripples, I was more or less thinking prevention, by the time Larvae is in the water, it is too late? how much time before they hatch etc.. (need a mosquito expert).

But good man, interested to hear what the other group may contribute, please let me know.

Cheers.


(Patrick) #13

Nitrite/Nitrate Sensing: We are placing sensors in water to measure nitrogen (and other things). [We're an island.] Is anyone using sensors to measure nitrites or nitrates in water? Love to know your set up— we'll use a MicroTech conduit.


(Brendan O'Gorman) #14

. @creatinghere good stuff. Is this aimed at aquatic life preservation or drinking water?


#15

@Brendan

"You don't think passive monitoring of the accelerometers in smart phones to identify and plot changing road conditions is a killer app (when holes and tires collide they produce a distinct motion and sound signature)?"

There is an app for that: http://www.streetbump.org


(Brendan O'Gorman) #16

@Lieutier Cool. Not surprised, had it mind for a while but didn't see a way to monetize , it's a generic kinda app (no not a 'killer' app :wink:

But looking at that site it seems to be not a very slick implementation, and doesn't seem to have gained traction.. also no android is a no no, still room for improvement, but free stuff.. so..