I’ve been increasingly hearing about local governments installing flood sensors as a way to alert the public about high-water situations. There seems to be an increasing number of major flooding events in the news, like Hurricane Hellene last year and the Guadalupe River floods earlier this summer.
But there are numerous smaller flooding events all of the time that result from heavy rains. In Appalachia, where I live, and in places like the Ozarks and the Rocky Mountains, floods can spring up quickly along roads after a rain event, often from upstream rain outside the flooded area. Flood sensors are badly needed in some parts of the country.
The danger from local flooding is that roads suddenly go underwater and people drive into the flooded area, particularly at night. I heard from one County that said that residents blindly drive into flooded areas every year. Historically, local governments blocked off roads after local flooding, but doing so is a delayed response after somebody notifies officials that a road is flooded.
The ideal flood sensor system uses a sensor that can detect rising water levels. One company that makes these sensors is HyFi. These sensors can detect rising water before a full flood situation occurs. The ideal flooding system also includes a way to alert the public. In my County, there are text alerts issued when roads flood. The ultimate safety solution is a system that blocks off roads, using technologies like dropping a crossing gate, like those used at railroad tracks. Flood alerts can also use electronic signboards that warn drivers of a flood ahead on a road.
The ideal flood sensor system also includes some kind of broadband connectivity. Local public safety staff want to know when a flood is underway to be able to alert the public. In urban areas with great cell coverage, flood sensors can be connected via cellular technology. But in rural areas, a lot of areas have poor or no cell coverage, and the best solution is to connect flood sensors to a fiber network.
Flood sensor systems are a great safety measure along roads that quickly and routinely flood. It’s a lot harder to develop a flood sensor system that works during catastrophic flooding. In Hurricane Helene, the flooding was so severe that entire roadways and even whole towns were washed away in the floodwaters. Early during the hurricane, the cellular networks went dark after winds affected microwave dishes, and when floods washed away the fiber backbone lines that served all of the ISPs and cell carriers in the region. In areas with the worst flooding, the electric substations were also put out of service. There is a lot of speculation that many who died in the hurricane didn’t get any notice that the floods were of historical proportions.
But catastrophic floods are rare, while smaller floods are routine in some places. Flood sensors, along with the broadband backup, ought to be on the wish list of any locale that routinely floods. A lot of the people who die in floods drive into the rushing waters – something that could be avoided with good flood detection and alert systems.
It seems a flood sensor doesn’t need a lot of bandwidth. Sensor number, timestamp, water level (or an indication that the water level has exceeded a pre-determined threshold) and a “periodic all is well heartbeat” . Unless one is wanting to monitor the levels from a long distance away, a land mobile radio based SCADA system should be enough….or am I missing something?
we run many thousands of sensors using lorawan, zigbee, some proprietary, amazon sidewalk etc. There is absolutely no need for broadband for thousands and thousands of sensors. You’re talking about data bursts of 10-20bytes.
For a reliable sensor network you need very basic connectivity and preferrably redundant connectivity on separate providers and technologies.
Our preference is to run lorawan sensors. Multiple gateways that can receive and handle duplicate entries connected to separate internet providers makes for high reliability, timely data at extremely low costs.
Even utility scale sensors need extremely little. We pass water data from a community through and they have an unnoticable amount of throughput. Their web browser to access the system uses thousands of times more data than the network does.
‘broadband’ is thousands and thousands of times more capacity than is needed.
I did not immediately think of LORA, but that certainly seems doable. I have not used it but SUSPECT it is largely line of sight due to the frequencies used. Does that work for stream sensors in the middle of nowhere?
Lorawan in the US is in 900Mhz. We’ve hit 120km line of site on a sensor. however, in dense trees is a few km. in a city getting through homes it’s a bit under 1km. can be extended with better antennas. We have some cistern sensors with 3′ antennas that get through some pine trees and then a few dozen km with ~90% of the chirps making it.
Lorawan is amazing technology! I need to play with it some day.
Harold
https://w6iwi.org
when you get properly into IoT, the nb-iot/cat-m prices are even lower. sub $1/sensor, and at scale even less if you are doing readings less often. And take it a step further and use a no-IP service like Verizon offers and it’s even less.
however, get into lorawan and use ttn or helium and your costs are 1/10 nb-iot/cat-m or less. You can take 1 minute readings on helium for that, if you’re doing once a day it’s 10,000 times cheaper if there’s coverage.
In a county where I lived, they had level sensor in all the creeks. These would randomly report the water level on a VHF frequency licensed to the county using ALOHA access (they would transmit their data at random times, and there would occasionally be a collision, but occasional missed data was not a problem). The water level sensors did not have receivers (which would have allowed for polling). They just had transmitters. This DID require infrastructure, though. I think they ran this through the animal control department’s repeater. They probably had some method of keeping the data bursts out of other receivers (perhaps disabling subaudible “PL” tone transmission by the repeater during the data burst).
An advantage to the use of cellular for this sort of thing, even though the amount of data is very low, is that someone else gets to pay for the infrastructure. For example, https://www.digikey.com/en/resources/iot-resource-center/iot-cellular-data-plans shows plans for as little as $3 per month.
Another low data rate system is the “smart meters” where, as I understand it, the meters form a mesh network that connects to a cellular gateway to get the data onto the internet and back to the utility. This works well when there is a high density of devices (the meters) since only one cellular interface is required for a lot of meters. For the stream water level system, though, I suspect the device density is pretty low, so the VHF/UHF system or cellular data is required.