Medical refrigerators store critically sensitive materials like vaccines, blood and clinical samples. Knowing if a refrigerator is under stress and about to fail is valuable information that allows stakeholders to react more quickly, saving money and loss of life-saving contents.
In this project we measure the health of the refrigeration system by using four digital temperature sensors placed at strategic points: compressor exit, condenser exit, case and ambient. A Raspberry Pi is used to read the sensors, filter and package the data into a secure payload and transmit to cloud services for analysis.
Just like an EKG is used to monitor a heart and look for anomalies, so we use time series analysis to look for higher than expected duty cycles, or out of the ordinary waveforms.
All hardware components are integrated using a Rapid Prototyping kit from Zymbit. Analytics are performed in real-time using InfluxDB and Grafana software components.