The present paper deals with the development of a Fault Diagnosis and Prognosis supervision module which can be applied to LED lighting systems for both industrial and domestic applications. Hardware redundancy is used for detecting and isolating a fault on the LED lamp (composed of a fixed number of LED strings). Residual signals are obtained as the difference between the currents circulating into two different LED strings piloted by the same input current. A signal-based approach is used for fault detection on the cooling fan subsystem, where the diagnosis system is based on the heat sink temperature. Fault detection is performed through an algorithm based on the trend of the temperature rate on the heat sink. Fault detection on the light, motion and temperature sensors is based on a straightforward evaluation of the sensor readings. In all these cases, a threshold policy has been used to trigger the faulty condition. Finally, the module integrates a model-based prognosis algorithm which is used to predict when LED maintenance should be performed. The proposed solution has been experimentally validated, showing the capability to correctly detect and isolate a fault occurring on the LED string, on the fan subsystem, or on the sensor set, and providing at the same time an estimation of the LED useful life.
A Fault Diagnosis and Prognosis LED Lighting System for Increasing Reliability in Energy Efficient Buildings
FREDDI, ALESSANDRO;
2013-01-01
Abstract
The present paper deals with the development of a Fault Diagnosis and Prognosis supervision module which can be applied to LED lighting systems for both industrial and domestic applications. Hardware redundancy is used for detecting and isolating a fault on the LED lamp (composed of a fixed number of LED strings). Residual signals are obtained as the difference between the currents circulating into two different LED strings piloted by the same input current. A signal-based approach is used for fault detection on the cooling fan subsystem, where the diagnosis system is based on the heat sink temperature. Fault detection is performed through an algorithm based on the trend of the temperature rate on the heat sink. Fault detection on the light, motion and temperature sensors is based on a straightforward evaluation of the sensor readings. In all these cases, a threshold policy has been used to trigger the faulty condition. Finally, the module integrates a model-based prognosis algorithm which is used to predict when LED maintenance should be performed. The proposed solution has been experimentally validated, showing the capability to correctly detect and isolate a fault occurring on the LED string, on the fan subsystem, or on the sensor set, and providing at the same time an estimation of the LED useful life.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
