Is there a need for the application of PHM to LED lighting systems? In recent years, LED systems have become extremely popular as an environmentally friendly alternative to both incandescent and fluorescent lights. As governmental restrictions on power usage and hazardous substances become tighter (to the point of even banning the usage of incandescent lamps in certain settings) , LED systems should continue to grow in popularity not only for personal usage and consumer electronics, but also for public settings as with street lights, in which reliability would be a major safety concern.
Unfortunately there are a number of barriers preventing LED systems from completely replacing incandescent and fluorescent lighting systems. The most obvious factor is the life cycle costs, which are ten to twenty times higher than that of equivalent fluorescent lights. These costs are distributed among premarket qualifying operations as well as postmarket maintenance and support service. The most prominent challenges today include:
These issues are compounded by the fact that LED degradation can only be predicted by simultaneously considering several factors such as ambient temperature, forward current, relative humidity, mechanical stress, and materials. And – even taking these factors into consideration – there are still severe inaccuracies involved in determining LED lifetime trends. These predictions tend to vary greatly with the sample size as well as test duration. As a result of these variations, LED lifetime durations published by manufacturers often have no physical meaning, resulting in costly maintenance fees and untimely failures.
A perfect example of this is the unexpected shutdown of LED lamps used for street lighting in Xiamen in 2008. The Municipal Government purchased these LEDs from a manufacturer guaranteeing 100000 hours of life (approximately 5 years of proper usage), but the lighting system began to fail after only two months of usage. Despite these types of failures, the market size of LED lighting systems in China has increased astronomically in recent years with plans to breach 80 billion RMB by 2015 in the Guangdong province alone. With so much investment (3.5 billion RMB over the last 4 years) into such a large market, there are clearly massive applications for a PHM system to improve LED reliability.
The goals of using a PHM approach for this problem are widespread, including:
PHM can accomplish these goals by utilizing condition monitoring and parameter analysis methods. Sensors can be designed to monitor variables such as temperature, luminous flux, chromaticity coordinates, voltage drops, and current leakage. These loads reflect degradation and future reliability. Analyzing load data makes it possible to not only more accurately predict remaining usage life, but also to take automated control actions (i.e. altering the environmental or operational conditions) in order to increase service lifetime. This can greatly improve maintenance effectiveness and also help avoid costs associated with redundant inspections. PHM also makes it possible to isolate critical parameters or signals associated with specific types of failures, allowing us to trace anomalies back to their root cause and improve upon future designs. In other words, PHM serves as a continuous picture of overall system health, making it easier to deal with and prevent aging problems associated with LED lighting systems.
How has the CityU PHM Centre begun to tackle PHM implementation? The Hong Kong Government has begun to consider using LED lighting to replace the existing 130000 high pressure sodium street lamps to improve energy efficiency. However they are concerned with LED lighting reliability and system health monitoring. Avoiding a situation similar to the failure in Xiamen is critical. The CityU PHM Center is currently developing a prognostics-based solution to build an LED lighting system for a pilot installation of LED street lights as well as for other Hong Kong and Mainland China industries.
To avoid unexpected or premature failures, the most pressing issues for LED system development are qualification and lifetime estimation. Currently lifetime estimations are made using accelerated life tests performed at high temperatures while monitoring luminosity. Lifetime estimates are then calculated using an Arrhenius model. This involves plotting luminosity versus time and then scaling the time down based on an acceleration factor. Unfortunately not only is the acceleration factor inaccurate, but the Arrhenius model does not represent all modes of failure for LED lights. A more complete model must consider temperature, forward driving current, relative humidity, mechanical stress, and materials. Keeping this in mind, the CityU PHM Centre is currently developing a more advanced life qualification tool that can predict the lifetime of LED systems during the design, development, and early production phases.
