The Research On Reliability Analysis And Prediction Of LED Street Lighting System

Optoelectronics industry is one of the most competitive in the internationalhigh-tech fields, where the development of light-emitting diodes have more worldattention. As LED performance continues to enhance and improve, application areas areexpanded. In the environment of development energy conservation, low-carbon mode,LED has much attention on the development of street lighting area, which is expected tobecome the fourth-generation illuminating sources to alternative traditional sources,become the preferred way of the future lighting. LED street lighting can be spread,depending on their reliability and lifetime prediction accuracy is improved. The studyon LED street lighting reliability is the premise and foundation to improve theirreliability. This thesis focuses LED street lighting system, using the theory of reliabilityanalysis to analyze and study the different modules, like LED modules, drive power,and cooling devices which can affect the reliability of LED street lighting.First, a new lifetime prediction method was proposed for LED this high cost,long-life products. Degradation trajectory model was established based on performancedegradation data. According to the model performance degradation value was predictedand combined with the failure threshold predict failure time, ie based on Particle SwarmOptimization algorithm optimize Wavelet Support Vector Regression method. Throughinstances study to verify, the accuracy of the established degradation trajectory modelhigher than usual exponential decay model.Then, the research based on LED driver power working principle and reliabilitytesting, combined with drive power analog circuits and Monte Carlo analysis, themethod of Particle Swarm Optimization algorithm optimize Support Vector Machinewas proposed to fault isolation and classification.Finally, based on these studies, LabVIEW software was applied to develop aspecifically software for LED reliability analysis. The purpose is to make the reliabilityanalysis of LED products in project more targeted, more convenient operation. As adedicated LED product reliability analysis software, the major function of the softwareis intuitive descripting the commonly reliability parameters of the LED products, suchas the average life, characteristic life, loss of efficiency and the average loss ofefficiency. Through examples verified analysis, the software has achieved good results in high power LED reliability analysis research work, and obtained National ComputerSoftware Copyright Registaration.