Towards Prognostics & Health Management in Lighting Applications



D. Schenkelaars W.D. van Driel M. Klompenhouwer I. Flinsenberg R. Duijve


Philips Lighting’s revenue is largely influenced by the change from component supplier to supplier of systems, solutions and services. Philips Lighting differentiates from competition by providing high quality and reliable products as we learned in our traditional lighting business and which we continue in our actual LED lighting business. Reliable products start with understanding the physics-of-failure by using accelerated test approaches such as (Highly) Accelerated Life Testing. A classical reliability approach is to use the results from these tests, verified by failure analysis, to obtain conservative bounds from the failure models, and predict failure rates on a system level. A next step beyond this classical approach is to use data analytics in our installed base to determine degraded performance. The data for this analysis can come either from live connections to ‘intelligent’ systems, or from actively retracted (working) products from the field. This allows us to move into the prognostics (PHM) regime where a detailed understanding of failure mechanisms, usage scenarios, technology and design come together.
Until recently costs for implementing PHM in Lighting products or systems was outranging possible cost benefits. Nowadays this is reversing rapidly by the exponential increasing impact of digitization and connectivity of the Lighting Systems. The impact is far beyond the impact on single products, but extends to an ever larger amount of connected systems. Continuously, more intelligent interfacing with the technical environment and with different kind of users is being built-in by using more and different kind of sensors, (wireless) communication, and different kind of interacting or interfacing devices. Especially in professional systems, where many years of use has to be warranted and where system size, cost and complexity are continuously increasing, PHM is required. Where online debugging and adding new features or functions is already common practice, PHM should provide tools to keep the system within its quality and reliability targets. In the presentation we show our road towards prognostics and demonstrate PHM work being done in different professional Lighting applications as Public Lighting, Office & Industry Lighting and Retail Lighting. While data analytic tools are still premature, first results are achieved and improvement tracks are being defined. We will conclude with our strategy and vision on how to embed cost-effective PHM into lighting applications.

How to Cite

Schenkelaars, D., Driel, W. van, Klompenhouwer, M., Flinsenberg, I., & Duijve, R. (2016). Towards Prognostics & Health Management in Lighting Applications. PHM Society European Conference, 3(1).
Abstract 43 | PDF Downloads 34



Lighting LED System Reliability Degradation Availability Data Analysis Maintenance Model

Breiman, Leo, (2001). "Random Forests". Machine Learning 45 (1): 5–32.doi:10.1023/ A:1010933404324.
Driel, W.D. van, Fan, X.J., Solid State Lighting Reliability: Components to Systems, ISBN 978-1-4614-3066- 7, 31 August 2012, Springer, 617 pages.
Nelson, Wayne B., Accelerated Testing: Statistical Models, Test Plans, and Data Analysis... ISBN: 978-0-471- 69736-7. 624 pages. September 2004.
IES, LM-80-08: Approved method for measuring maintenance of Led light sources.
Driel, W.D. van, Schuld, M., Jacobs B., Commissaris, F., van der Eyden, J., Hamon B., Lumen maintenance predictions for LED packages, Microelectronics Reliability (2016), j.microrel.2016.03.018
Lima, S. de, Trivedi, H. G., Driel, W.D. van, Methods and apparatus for end-of-life estimation of solid state lighting fixtures, IP WO2012/156857, 2012.
Technical Papers