Streams and rivers are crucial components of many ecosystems and human activities. Today, the USGS maintains the largest collection of stream sensing devices. They have about 10,000 sensing devices across the nation. Unfortunately, this does not cover millions of American homes susceptible to flooding from nearby streams and rivers. Additionally, these sensing stations are designed for accuracy. This means that installing one has a high cost of infrastructure and can cost thousands of dollars.
Stream data does not have to be limited to gauge and flow rates. Many residential drinking supplies come from water that originates as runoff. Impurities like road salt and fertilizer can sometimes find their way into these drinking supplies. By monitoring temperature and conductivity in these streams, you can help predict ecological events like algal blooms. And, if the network of sensors is large enough, you can pinpoint the origin of the impurities.
By creating a low-cost, distributed sensor network, we hope to address both of these problems. The availability of large volumes of high-quality sensor data has the potential to improve applications such as weather forecasting and ecosystem health monitoring.
(A) A Particle Electron microcontroller is responsible for coordinating all the sensors present on the device and using a cellular data connection to transmit this to the central server.
(B) A 3G antenna is mounted internally away from other electronics to improve reliability.
(C) An ultrasonic distance sensor faces the surface of the water measuring fluctuations in the height of water flowing through the stream or river. The raw sensor data can then be used as a measure of relative flow rate.
(D) A thermistor mounted under the surface of the water measures the temperature of water flowing through the stream or river.
(E) AA batteries were chosen as the device’s power source for their high energy capacity, low leakage current, and resilience to extreme temperature. Cells are wired in series in sets of 3. This comes out to 4.5V with 18 total cells.
(F) A 3D printed enclosure is waterproof and designed to be easy to install at the field.
(G) An Arduino microcontroller uses a MOSFET to switch power to the main Boron microcontroller. This reduces the device’s power consumption and extends battery life significantly.
(H) Conductivity is measured by passing a small current through two aluminum plates. When placed under the water, it can monitor relative conductivity.
Bill of Materials
3D printed body
Particle Electron 3G
Ultrasonic Distance Sensor
AA Batteries x 15
Passives, wire, hardware
The software powering the project consists of two main components: the embedded software running on the actual sensor, and the software powering the web server which receives and stores the data from the sensor and provides historical data to users.
The embedded software is designed to consume as little power as possible to extend battery life. The Particle Electron stays in a deep sleep state in between readings which are taken each hour. When the Particle wakes, it takes sensor readings at the same time that it attempts to connect to cellular data. This is so that when it does eventually establish a connection, it can immediately publish the data and return to a sleep state. The device has a software watchdog running as well. If the sensor is on for more than 5 minutes without connecting, it returns to sleep and tries again later.
The web server serves two main functions: to receive data from the individual sensor units, and to make this data available to users. To submit data to the server, a sensor unit submits an HTTP POST request consisting of a JSON string which contains the current sensor readings. The server parses this string and adds the data to...Read more »