Intelligent Prediction on Heat Dissipation Capability of the High-power LED Street Lamp Based on Fusion Function Chain Neural Network

Abstract

Growth of the power for LED and temperature of junction will lead to rapid decay of LED. For this characteristic, the heat pipe technology was used in the radiator. We manufacture a new radiator improving cooling efficiency, test it at 25 ◦ C and use EMD method to remove the noise. Moreover, the forecasting model can be obtained after a fitting formula about the temperature inside the lampshade at different LED working hours is established by using the function chain neural network. The result shows that the temperature inside lameshade is 54 ◦ C after the new high power street lamp has worked for ten hours, which illustrates the new high power street lamp has good heat dissipation capability.