Remote monitoring of LED lighting system performance

The concept of connected lighting systems using LED lighting for the creation of intelligent buildings is becoming attractive to building owners and managers. In this application, the two most important parameters include power demand and the remaining useful life of the LED fixtures. The first enables energy-efficient buildings and the second helps building managers schedule maintenance services. The failure of an LED lighting system can be parametric (such as lumen depreciation) or catastrophic (such as complete cessation of light). Catastrophic failures in LED lighting systems can create serious consequences in safety critical and emergency applications. Therefore, both failure mechanisms must be considered and the shorter of the two must be used as the failure time. Furthermore, because of significant variation between the useful lives of similar products, it is difficult to accurately predict the life of LED systems. Real-time data gathering and analysis of key operating parameters of LED systems can enable the accurate estimation of the useful life of a lighting system. This paper demonstrates the use of a data-driven method (Euclidean distance) to monitor the performance of an LED lighting system and predict its time to failure.