Study On Online-Test And Fast Lifetime Evaluation Of LED Lamps In Accelerated Aging Test

With the advantages of long lifetime, energy conservation and environmental protection, LED as the new energy source of fourth generation has been widely used in various lighting fields. While its advantages in energy conservation and environmental protection have been widely admitted, its long lifetime and good reliability are still being questioned. Therefore, many scholars have devoted to the related study. At present, the lifetime prediction for LED lamp is mainly based on the standards of Energy Star proposed by Illuminating Engineering Society. The recommended test time is at least 6000 hours, which is not only expensive to perform but also time-consuming that causes an obstacle for new products updated.Therefore, the lifetime evaluation for LED products in a short period of time becomes researchers' attention in the field of LED industry of the world.1. The issue of how to choose the highest temperature stress level is studied,which can shorten the time of accelerated aging test of LED lamps under the condition of failure mechanism unchanged. Firstly, temperature step-stress test is applied to LED lamps, and the parameters of luminous flux, correlated color temperature, color shift and junction temperature of lamps are acquired at each step of the test. Secondly, the relationship between the parameter and the temperature stress is analyzed. There is a catastrophe point in the variation of the parameter with respect to the temperature stress, a sudden rising or falling as temperature stress increasing. The temperature stress corresponding to the catastrophe point is taken asthe limit of stress level for the accelerated aging of LED lamp.2. The constant stress accelerated aging online test is studied and the online test platform is built. Progress is made in the following aspects.1) Based on the exponential decay law and the 5-order moving average error,the evaluation errors of accelerated lifetime of LED lamps are obtained under different conditions of the cut-off time. The optimal cut-off time is then obtained under correspondent temperature.2) The approximate method, analytical method and two-stage method are respectively used in the data processing. The unknown parameters of reliability distribution(Weibull distribution and lognormal distribution) are then obtained. At the same time, AIC information criterion is used to evaluate the models and the results show that the two-stage method is the best model.3) The two-stage method is used to calculate the accelerated lifetime based on the data of online test at temperature of 60? and 80?. The Arrhenius model is then used to deduce the lifetime at room temperature of 25?. In a comparison, the aging test of lamps from same batch is conducted based on Energy Star standard and the results show that the online test has a satisfactory reliability.4) During accelerated aging test at temperature of 80 ?, the chromaticity parameters(correlated color temperature(CCT) and color shift) of LED lamps are measured and analyzed by both online test and offline test. The color shift and the increment amount of CCT given by online test are about 20% higher than that given by offline test. In particular, the percentage change of CCT by online test is only6.8% higher than that by offline test. The t test and t' test are respectively used to do statistical analysis for the data of CCT and color shift. It is shown that the above differences are of statistical significance.3. Based on the concept of constant stress accelerated aging test, the conversion model and the steps of step-stress accelerated aging test are studied.1) The Energy Star standard is used to verify two conversion models in data processing of step-stress accelerated aging test of LED lamps. And the results showthat the model of equal cumulative amount of degradation is more suitable than the model of equal cumulative failure probability.2) Four-step-down accelerated aging test of a type of LED lamps is done to determine the stresses in two-step-down aging test with the criteria of minimum standard deviation of reliability and minimum standard deviation of relative lifetime of samples, and then the activation energy is calculated.4. The automatic and fast test platform of online step-down aging test of LED lamps including hardware equipments and software of lifetime prediction is studied.Firstly, hardware equipments including turnable of lamp shelf, picker mechanism and motion control system are developed. Secondly, the above optimal algorithms are integrated as software to achieve automatic lifetime evaluation of LED lamps. In order to verify the accuracy of online step-stress accelerated test, the aging test of lamps from same batch is also conducted based on Energy Star standard. It is shown that the difference of liftetime prediction between the two methods is 3.8%,4.1%and 4.2% when the failure probability is 63.2%, 10% and 1%, respectively.In this thesis, the constant stress and step-stress accelerated aging tests for certain type of LED lamps are conducted, and progress is made. The conclusions of this research may play a guiding role in the accelerated aging test of other types of LED lamps.