Study On Prediction Model Of High Power LED’s Junction Temperature

With the tightness of energy supply, the LED is widely used in various industries for its energy efficiency and long life. Although the LED luminous efficiency is improving, about 80% to 90% of the energy is distributed in the form of heat. The heat causes the problem of temperature rising and the waste of energy, which affect the development of LED industry seriously and resulted in LED’s aging or burned. It is necessary for real-time monitoring of the LED junction temperature to achieve high reliability of LED.Based on the LED principles of electroluminescence and the photoelectric theory, the influence of the LED input optical parameters of power, junction temperature, heat set and luminous flux to the life of LED are studied. The prediction model of the LED’s life is established and the key factor is indicated which is junction temperature.The 120 W high power LED with double inlet and outlet jet system is chose as the object of study. According to the law of conservation of energy and the LED heat transfer, the thermal conductivity model is obtained. Use the step response of area identification method to obtain the power-junction temperature transfer function model of the LED, and the transfer function model is simplified by using the closed loop dominant apices. Comparing of the step response model making use of the measured data is to verify the validity of the transfer function model.In order to get a more accurate predictive value of the junction temperature, the BP neural network is used and the prediction model of high-power LED junction temperature is established. The LM algorithm is used to improve the BP algorithm. The convergence of the improved model is faster, but the generalization ability of BP neural network and the shortcoming of falling into local minima easily are not improved. In order to improve these defects of BP neural network, the genetic algorithm is applied to optimize the prediction model of LM-BP junction temperature. Finally, the prediction model of high power LED junction temperature is gotten, and the validity and accuracy of the model are confirmed.The LED junction temperature changes can be obtained by using the junction temperature prediction model. It is good to keep the junction temperature of the LED in a suitable condition and achieve active control of the LED junction temperature which prevents the LED junction temperature from too high. Besides, the model can maximize the life of the LED, and ensure the stable and reliable of the LED lighting system.