A generic changepoint detection algorithm monitoring sensor data to better predict boiler malfunctions
bCheck is on a mission to make boilers smarter, mainly targeting real estate management companies. They deliver solutions to predict boiler malfunctions or breakdowns, with focus on minimizing costs on maintenance or replacements.
bCheck thrives to optimize energy consumption and to protect households from boiler breakdowns in winter periods.
How can we predict breakdowns on boilers?
bCheck has developed its own sensor devices that can be attached to boilers, helping them in collecting a significant amount of data on parameters like temperature, throughput and vibrations. Having this data at its disposal, bCheck wanted to leverage the power of Machine Learning to see if they were able to predict whenever a boiler was heading to breakdown.
Using Infofarm’s expertise on Artificial Intelligence/Machine Learning algorithms, bCheck gained a valuable partner on their quest to predict boiler breakdowns. This however did not turn out to be an easy task. In order to make a substantiated prediction, historical data of boiler breakdowns was needed, but not available yet.
Therefore, we started to look at the real-time behavior of a sensor and its data, in order to detect anomalies or changes in that behavior. These flags could then later be used to detect malfunctions and to send automated alerts to maintenance staff.
SMART MONITORING TOOL
An interactive monitoring application using a generic Changepoint Detection algorithm to flag anomalies or changes in sensor behavior.
Intelligent messaging system to optimize and automated maintenance tasks and planning.