Cooling Resear And Life Prediction Of LED Headlamp Chip

With the rapid development of the automobile industry,as the eyes of the automobile,the automobile lights is great of importance.Modern automobile light sources mainly choose energy-saving,environmental protection,long life LED(light emitting diode).But because of its high power,small size,most of the energy convert into heat,result in LED chip temperature is high relatively,heat dissipation problem has become the most important factor affecting the service life of LED.Therefore,to solve the problem of heat dissipation and predict its service life will inevitably promote the development of LED,and bring great economic benefits.In order to confirm the real reliability in simulation,this paper use 3D modeling software Solidworks to draw a 1:1 model according to the real radiator and car lamp LED chip,and the material parameters of each part of the device are exactly the same as the actual situation.This paper,by using the finite element analysis software ANSYS Workbench to build3 D radiator-chip model for steady state thermal simulation analysis,At the same time,use the thermocouple test principle to test the node temperature of the chip under normal operation,comparing simulation and experimental results,proved that the simulation model is accurate relatively and ANSYS Workbench used in high-power LED headlights thermal analysis is feasibility.Using ANSYS Icepak,Steady-State Thermal module for Thermal analysis of the mathematical models of chip,from the shape of a coating,built-in Thermal column cross-sectional area,coating internal hole interval size,the density of radiator fin,research the four aspects influence on heat dissipation rate.find the connection between highest temperature of the chip and the four aspects,provide certain theoretical support for improving the heat dissipation.The life of LED car lamp chip is affected by multiple factors,the most important of which is current stress and temperature stress.Based on the sample performance index,choose currents of 1200 mA,1400mA and 1600 mA as the acceleration stress level.In order to shorten the test time,the light flux attenuation to the initial 70% was deemed to invalid.When processing test data,used Weibull distribution to describe the life distribution of LED chips of automobile headlights,And use the least square method to make statistics.Finally,accelerated life model(inverse power ratio model)was selected to calculate the life of LED headlights under normal stress level.The LED chip current-life BP neural network was constructed to take the loading current,the initial luminous flux,the luminous flux after loading as the input,the chip life under acceleration stress and the life under rated stress as the network model of the output of the neural network.Taking the data from the constant current loading experiment as the sample of network training,the network(number of hidden layer nodes,training times)was adjusted by adjusting the corresponding parameters in the network to make the prediction accuracy within the set range.Finally,a group of untrained samples were input into the trained network model,and the predicted life was compared with the experimental life.It was found that the error between the predicted life and the real life of the neural network was between 2% and 4%,indicating this network could be applied to the life prediction of LED chips.In short,this paper mainly studies the improvement of heat dissipation performance of car lamp LED chip through finite element simulation.The working life of LED chips was predicted by the combination of constant current accelerated life experiment and BP neural network.This study provides a theoretical reference for optimizing the thermal design of automobile LED lights and an important guidance for establishing the life prediction model of LED chips.