Acoustic resonance detection using statistical methods of voltage envelope characterization in metal halide lamps - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue IEEE Transactions on Industry Applications Année : 2017

Acoustic resonance detection using statistical methods of voltage envelope characterization in metal halide lamps

Résumé

Acoustic resonance (AR) phenomenon occurs in metal halide lamps and can cause light flicker, lamp arc bending and rotation, lamp extinction, and in the worst case, arc tube explosion. This study takes place in the context of developing electronic ballasts with robust AR detection and avoidance mechanisms. To this end, a lock-in amplifier is used to measure and characterize lamp voltage root mean square (rms) short-term fluctuations. Statistical criteria based on the standard deviation of this rms value are proposed to assess AR presence and classify its severity. A set of metal halide lamps from different manufacturers and with different powers were tested. The average electrical power and AR level are controlled by adjusting the lamp operating frequency of high-frequency electronic ballast. The proposed criteria enable classifying healthy (without AR) and faulty (with AR) cases based upon either a two-dimensional plane or a boxplot. Regardless of lamp operating power, the results from this study show that the voltage rms variations and the defined criteria are significantly correlated with AR level.
Fichier principal
Vignette du fichier
Lei_21457.pdf (619.1 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02001394 , version 1 (31-01-2019)

Identifiants

Citer

Fang Lei, Pascal Dupuis, Olivier Durrieu, Georges Zissis, Pascal Maussion. Acoustic resonance detection using statistical methods of voltage envelope characterization in metal halide lamps. IEEE Transactions on Industry Applications, 2017, 53 (6), pp.5988 -5996. ⟨10.1109/TIA.2017.2742978⟩. ⟨hal-02001394⟩
45 Consultations
77 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More