| The Solar Simulator is an instrument that simulates the characteristics of solar radiation by using an artificially controllable light source and is widely used in the photovoltaic industry.LEDs have gradually replaced xenon lamps and halogen lamps as the main source of solar simulators due to their full spectrum and controllability.In order to achieve higher optical power density in a limited space,solar light simulators must contain multiple LEDs working in high power state,this will cause serious thermal coupling effect between LEDs.Thermal coupling effect is the influence of each heat source on the temperature distribution of each other,this will result in different temperature rise of LED junction temperature,it is reflected in the uneven distribution of LED junction temperature,that is the existence of heat island phenomenon.The photoelectric thermal theory describes the mutual coupling relationship between the LED photoelectric parameters.It shows that as the temperature rises,it will affect the optical properties such as luminous flux and peak wavelength,Therefore,it is very important to effectively reduce the intensity of thermal coupling effect,and the thermal coupling effect is affected by many factors such as ambient temperature and light source structure,and each factor is coupled with each other,which brings great challenges to the study of thermal coupling effect.In this paper,the solar light simulator multi-LED light source is taken as the research object.In order to solve the problem of LED thermal coupling effect,the neural network model about input parameter wavelength ?,drive current If and output parameter luminous flux ?,electric power Pe and thermal power Pl is established.And analyzed the LED spacing d,thermal conductivity of heat-dissipating substrate,thickness h of heat-dissipating substrate and so on.And reduce the thermal coupling effect by changing the spatial arrangement of the LEDs.The specific contents are as follows:(1)Photoelectric thermal theory can approximately describe the mutual coupling relationship of single LED photoelectric thermal parameters.When the type of LED is changed,it is necessary to change the parameters such as Rjc,ke,kv in the theory,and the model is high-order expression with high computational complexity,which will undoubtedly bring limitations to the application of photoelectric thermal theory.In this paper,the BP neural network method using momentum term is proposed to express the mutual coupling relationship between photoelectric and thermal.The modeling data is obtained by building the LED characteristic test platform.Making comparison the model prediction results with the experimental results,it shows that the model has higher accuracy and speed when the model is used to analyze the photoelectric thermal coupling.(2)The neural network model of photoelectric thermal is used,and the thermal power of LED under different working conditions is used as the boundary condition for analyzing the thermal coupling effect.In this paper,being based on the equivalent model of LED circuit,using finite element method and thermal resistance matrix theory to analyze the influence of LED spacing,thermal conductivity of heat sink substrate and thickness of heat dissipation substrate on thermal coupling effect,and establish various factors and thermal coupling effect.The relationship model between the different factors and LED junction temperature rise which was caused by thermal coupling effect was established,which provides a convenient condition for analyzing the thermal coupling effect.(3)In the specific design of the light source,considering the thermal conductivity of the material,processing technology and other factors,the thermal conductivity of the heat-dissipating substrate and the thickness of the heat-dissipating substrate are limited,so the LED spacing is the main factor which could factor the thermal coupling effect.In this paper,being based on the model of LED junction temperature and temperature rise which was caused by thermal coupling effect,in the two array structures of rectangular array and circular array,it is proposed to arrange LED according to the law of variation of equal difference,Fibonacci and so on.And extend it to the function model of logarithm,power,distortion and so on,and using the variance as the fitness function to apply the genetic algorithm to optimize the parameters in the function model,and after being optimized,the LED junction temperature of the three functions.The variances are changed to 62%,20%,and 15%respectively.It can be seen that the multi-LED light source adopts the function model arrangement could obtain a more uniform junction temperature distribution.The reasonable space arrangement ensures that the LED junction temperature distribution is uniform and reduced.It is also an effective strategy for the strength of thermal coupling effect.In this paper,the BP neural network model is established to provide boundary conditions for the analysis of thermal coupling between multiple LEDs.The influence of multiple factors on the thermal coupling effect is studied.The LED spacing is selected as the optimization parameter of the layout,and the different layouts are compared.The temperature distribution shows that the variance of the LED junction temperature of the distortion function is 15%,which effectively reduces the thermal coupling between the multiple LEDs and improves the temperature distribution uniformity of the LED light source. |
PDF Full Text Request