Water plays a big role in supporting our communities. Without water there would be no local business or industry. Fire-fighting, municipal parks, and public swimming pools all need lots of water. An array of pipes, canals, and pumping stations managed by our public water systems are needed to bring a reliable supply of water to our taps each day.
In the US, we are lucky to have easy access to some of the safest treated water in the world—just by turning on the tap. We wake up in the morning, take a shower, brush our teeth, grab a cup of coffee, and head out for the day. Water is an important part of our daily lives and we use it for a wide variety of purposes, but do we really understand how much we use?
The average American family uses more than 300 gallons of water per day at home. Roughly 70 percent of this use occurs indoors. An easy to way to understand individual water use is to look at your water bill—not just the amount due, but how much water you used.
What it does
Our project does just that and provides an easy-to-use all-in-one water consumption dashboard. Our contribution provides both an IoT hardware solution to place on water outlets around the house (sink faucets, showers, bathtubs, laundry units, toilets, etc.) as well as a software solution to view water usage and compete with friends.
The hardware solution is a small IoT device that attaches easily onto pipes attached to water outlets. These small devices connect to a master hub within the household logging water flow through the pipes and computes how much water is used throughout the day. It can either connect to a water hose (toilets, laundry units), thread in a shower head or can be dropped into the pipe of a sink. When dropped into the pipe of a sink it can help prevent water waste going down the pipes and act as an obstruction. Inside is a mechanical water flow meter using a turbine that rotates with a magnet on it and is monitored by the hall effect and will be logged into the gateway and then sent into the server and logged into the dashboard.
The software solution is a dashboard that shows charts of water consumption by device (sink, shower, laundry), which is unique to each household and can be tracked by logging into an account. In addition, we have a ML Autoregressive model that uses a LSTM to predict the weather hourly for the next 7 days for the following features: temperature, cloud amount, precipitation, and humidity. This data will be compounded with the water consumption per device to predict how much water usage will be used training on the weather data. We have shown preliminary results with some sample data and weather and with more data can train a more accurate model.
Lastly, in the web interface dashboard is a competition page that provides an educational game for water and wastewater learning. This provides a friendly competition for households to compete with one another in using the least amount of water per month and educates the user along the way on ways to reduce water usage.
Challenges we ran into
One of our biggest challenges was getting data to transmit wirelessly from the Arduino on the water flow module to the gateway efficiently. This was a large feat to have a small water flow sensor connect to an IoT gateway wirelessly but we were able to achieve it and it worked better than we could have hoped.
Another huge challenge was aggregating all the usages (water amount used) from each sensor into our server and displaying the data effectively in intuitive charts. This took a lot of debugging but we developed robust REST API endpoints that enable easy data logging and data visualization.
Accomplishments that we're proud of
Our greatest accomplishment is the overall UI and how native the live updating looks. We think that our UI helps educate the user with effective visuals and graphics that make it easy to understand and comprehend how they utilize their water.
What's next for Watery doing in Philly
We look forward to enhancing the game aspect of our project and introduce leaderboards that compare all users with filters for best in nation, best in state, and best in local area. Additionally, we are excited to see the improvements to our prediction of water usage and weather with our future weather predictions.