Unsupervised Fault Detection of Reefer Containers: A Moments-Based SDP Approach

Abstract

Refrigeration containers are being used like never before. The growing desire for exotic goods in much of the western world, puts an ever growing strain on the refrigeration systems that are used for transport. During one trip, a refrigeration container might experience temperatures ranging from −25 to 40 degrees celcius, whilst being tossed around on the deck of the freighter. This work, develops a method to detect errors aboard these refrigeration containers, that can be used to lower the maintenance costs. Furthermore, a prediction of the “time-to-failure” is presented, which enables the refrigeration container owners to decide whether or not a refrigeration container can endure a trip.