Outlier Node
Predictive Maintenance Sensor for Heavy Machinery
Built a predictive maintenance sensor for heavy machinery for the Student Innovator of the Year competition at BYU. Built a full machine learning model pipeline including cleaning mass amounts of vibration data, performing feature engineering, and choosing and training a Random Forest Classifier model to put onto our sensor.
Even though we weren't a finalist, we had a blast doing this competition and got some really good leads from judges who were interested in our idea.
Our first version attached to an HVAC fan motor
This was our first working prototype for Outlier Node. We used this for data collection, and to test our machine learning model.
I had to learn two different CAD softwares for this project. Tinkercad was used for the first version, which was a very simple and easy software. For the second version, I learned Fusion 360, which is an industry standard for CAD design.
We spent four hours at the SIOY showcase presenting our final project to judges and students.
We spent three hours presenting our first version of our project to mentors, who gave us advice and feedback on our idea and business model.
At our final showcase, we displayed all of our project iterations and prototypes to show our complete engineering process.