针对朗伯型发光特性的LED光源, 采用田口方法和传统遗传算法过程相融合的混合田口基因算法(HTGA)对用于光束整形的TIR透镜进行结构优化设计, 以期在特定目标屏上获得均匀照明效果。此算法把透镜结构参数组合作为染色体, 以目标屏上所获得的照明效果为基础构建评价函数, 最终优化获得TIR透镜结构参数。整个过程在Matlab和Tracepro共同搭建的平台上进行。仿真结果表明, HTGA算法充分利用田口方法的参数设计特性, 提高了传统遗传算法优越性, 最终优化所得TIR透镜在实现光能利用率60%的情况下, 使照明均匀度达到88%, 有效改善目标屏上的照明均匀性。

On the basis of the Lambertian light emission pattern of LED, Hybird Taguchi-Genetic Algorithm (HTGA), the combination of the Taguchi method and the process of traditional Genetic Algorithm, is adopted to make optimal design of the TIR lenses used as beam shaper, the purpose is to meet the uniform illumination requirement of specific target surfaces. This algorithm uses the construal parameters as chromosome and constructs the evaluation function based on the illumination result of specific target surface, by which the structural parameters of TIR lens can be achieved. The whole process is performed on the platform based on the Matlab and TracePro. Results show that HTGA algorithms, along with parameter design features of the Taguchi method during the entire optimization process, can improve the performance of the traditional GA algorithm. With the light efficiency reaching 60%, the optimized TIR lens can efficiently improve the illumination uniformity on the target surface to 88%.