LED光源等效模型建立及均勻照度色度透鏡設(shè)計(jì)
發(fā)布時(shí)間:2018-08-31 20:42
【摘要】:發(fā)光二極管(Light Emitting Diode,簡稱LED)具有節(jié)能環(huán)保、體積小、壽命長等優(yōu)點(diǎn),廣泛應(yīng)用于各種照明如路燈、閱讀燈及室內(nèi)照明等。LED直接輸出光的光強(qiáng)分布近似朗伯分布,如果直接照明在目標(biāo)面上將形成中心區(qū)域照度值比較高而邊緣區(qū)域照度值較低的照度分布,這種照度分布是不適合用于照明。為了提高LED照明性能,設(shè)計(jì)照明光學(xué)系統(tǒng)改變LED的光分布具有重要的意義。對于設(shè)計(jì)照明光學(xué)系統(tǒng)而言,首要問題是建立一個(gè)光源模型。光源模型建立的是否精確,對設(shè)計(jì)照明光學(xué)系統(tǒng)非常重要。本論文中,我們采用了預(yù)測算法和全局優(yōu)化算法的組合方法,可以實(shí)現(xiàn)自動(dòng)尋找LED光源模型的等效參數(shù),從而建立LED等效光源。這種方法不僅可以建立精確的單色光源模型,而且可以建立光色復(fù)合的光源模型,這種復(fù)合模型可以等效白光LED光源。本文圍繞著LED光源等效模型的建立以及基于這種光源模型設(shè)計(jì)了自由曲面透鏡使LED的輸出光在照明面上形成了均勻的照度和色度分布,論文主要成果如下: 提出了一種建立單色LED光源等效光學(xué)模型的新方法,通過大量的模擬數(shù)據(jù),使用BP神經(jīng)網(wǎng)絡(luò)算法建立LED光源等效參數(shù)與光強(qiáng)分布之間的隱性函數(shù)關(guān)系,構(gòu)建等效LED光源模型的光強(qiáng)分布與實(shí)際LED光強(qiáng)分布相似度的評價(jià)函數(shù),評價(jià)函數(shù)是以LED等效光源模型的參數(shù)為自變量的。評價(jià)函數(shù)值越小,光源模型的光強(qiáng)分布與實(shí)際光強(qiáng)分布越相似,我們通過模擬退火算法來尋找最佳的等效參數(shù),使評價(jià)函數(shù)值達(dá)到最小。通過這種方法建立OSRAM兩款LED的光源等效模型,得到模型的光強(qiáng)分布和實(shí)際LED的光強(qiáng)分布之間的歸一化互相關(guān)系數(shù)NCC均在99%以上,證實(shí)了該方法的可行性。 分析了熒光粉層的濃度、厚度等參數(shù)對LED的光強(qiáng)分布和空間色溫分布的影響。使用神經(jīng)網(wǎng)絡(luò)和模擬退火組合的優(yōu)化算法,可以精確調(diào)控?zé)晒夥蹍?shù),使其輸出光的空間色溫分布與預(yù)定的空間色溫分布接近。以瑞豐光電的3535芯片為例,通過這種方法對熒光粉參數(shù)的調(diào)控,得到的熒光粉等效模型與該芯片空間色溫分布的相似度非常高,NCC高達(dá)99.1%。使用相同的方法,建立光色復(fù)合模光源。光色復(fù)合光源的評價(jià)指標(biāo)有兩項(xiàng):光強(qiáng)分布和空間色溫分布。因此要建立光色復(fù)合模型光源必須使這兩個(gè)指標(biāo)都與實(shí)際LED光源接近。對瑞豐3535芯片的光源進(jìn)行等效建模,建立的等效光色復(fù)合模型與實(shí)際芯片的光強(qiáng)NCC為97.5%,空間色溫分布NCC=97.5%,可知所建立的光色復(fù)合模型可以與實(shí)際LED光色分布有很高的相似度,該方法可以較好的模擬白光LED光源。 基于建立的等效白光LED光源,我們設(shè)計(jì)了一個(gè)透鏡使白光LED光源產(chǎn)生的光能夠在照明目標(biāo)面上產(chǎn)生均勻的照度和色度分布。該透鏡是一個(gè)自由曲面透鏡,透鏡內(nèi)表面是一個(gè)球面,,外表面是一個(gè)自由曲面,由于透鏡的結(jié)構(gòu)是旋轉(zhuǎn)對稱的,所以設(shè)計(jì)的時(shí)候我們可以只考慮自由曲面的一個(gè)輪廓進(jìn)行設(shè)計(jì),自由曲面的輪廓曲線采用了3次樣條曲線。采取了手動(dòng)調(diào)整初始結(jié)構(gòu),使其照度和空間色溫分布比較接近理想的照度分布和空間色溫分布。初始結(jié)構(gòu)確定以后,通過TracePro軟件的交互式優(yōu)化功能作進(jìn)一步的優(yōu)化,均勻度分別為93%和91%,與優(yōu)化之前的透鏡產(chǎn)生的照度和空間色溫分布相比均勻度分別提高了17%和8%。
[Abstract]:Light Emitting Diode (LED) has the advantages of energy saving, environmental protection, small size and long life. It is widely used in various lighting such as street lamps, reading lamps and indoor lighting. In order to improve the illumination performance of LED, it is important to design illumination optical system to change the light distribution of LED. For the design of illumination optical system, the first problem is to establish a light source model. In this paper, we adopt a combination of predictive algorithm and global optimization algorithm to automatically find the equivalent parameters of the LED light source model and establish the LED equivalent light source. In this paper, a free-form lens is designed based on the equivalent model of the LED light source and the free-form lens to form a uniform illumination and chromaticity distribution on the illumination surface.
A new method to establish the equivalent optical model of monochromatic LED light source is proposed. Based on a large number of simulation data, the implicit function relationship between the equivalent parameters of LED light source and the intensity distribution is established by BP neural network algorithm. The evaluation function of the similarity between the equivalent LED light source model and the actual LED light intensity distribution is constructed. The parameters of the LED equivalent light source model are taken as independent variables.The smaller the evaluation function value is,the more similar the light intensity distribution of the light source model is to the actual light intensity distribution.We use simulated annealing algorithm to find the best equivalent parameters to minimize the evaluation function value.By this method,we establish the light equivalent model of OSRAM two LEDs and get the light intensity distribution of the model. The normalized cross-correlation coefficient NCC between the intensity distribution and the actual LED intensity distribution is above 99%, which confirms the feasibility of this method.
The effects of phosphor layer concentration and thickness on the light intensity distribution and spatial color temperature distribution of LED are analyzed.The parameters of phosphor can be adjusted accurately by using the optimization algorithm of neural network and simulated annealing,so that the spatial color temperature distribution of the output light is close to the predetermined spatial color temperature distribution.Taking Ruifeng photoelectric chip 3535 as an example,the system is on. The equivalent model of phosphor obtained by this method is very similar to the chip's spatial color temperature distribution, and the NCC is as high as 99.1%. Using the same method, a photochromic composite mode light source is established. The light source of Ruifeng 3535 chip is modeled equivalently. The NCC of the equivalent light-color composite model and the actual chip is 97.5% and the space color temperature distribution is 97.5%. It is shown that the light-color composite model can be very similar to the actual LED light-color distribution. The method can simulate white LED light source better.
Based on the equivalent white LED light source, a lens is designed to make the light produced by the white LED light source produce uniform illumination and chromaticity distribution on the illumination target surface. So we can only consider one contour of the free-form surface when we design. The contour curve of the free-form surface adopts cubic spline curve. The initial structure is adjusted manually to make the illumination and spatial color temperature distribution close to the ideal illumination distribution and spatial color temperature distribution. The interactive optimization function was further optimized. The uniformity was 93% and 91% respectively. The uniformity was increased by 17% and 8% respectively compared with the illumination and spatial color temperature distribution produced by the lens before optimization.
【學(xué)位授予單位】:江南大學(xué)
【學(xué)位級別】:碩士
【學(xué)位授予年份】:2014
【分類號(hào)】:TM923.34
本文編號(hào):2216102
[Abstract]:Light Emitting Diode (LED) has the advantages of energy saving, environmental protection, small size and long life. It is widely used in various lighting such as street lamps, reading lamps and indoor lighting. In order to improve the illumination performance of LED, it is important to design illumination optical system to change the light distribution of LED. For the design of illumination optical system, the first problem is to establish a light source model. In this paper, we adopt a combination of predictive algorithm and global optimization algorithm to automatically find the equivalent parameters of the LED light source model and establish the LED equivalent light source. In this paper, a free-form lens is designed based on the equivalent model of the LED light source and the free-form lens to form a uniform illumination and chromaticity distribution on the illumination surface.
A new method to establish the equivalent optical model of monochromatic LED light source is proposed. Based on a large number of simulation data, the implicit function relationship between the equivalent parameters of LED light source and the intensity distribution is established by BP neural network algorithm. The evaluation function of the similarity between the equivalent LED light source model and the actual LED light intensity distribution is constructed. The parameters of the LED equivalent light source model are taken as independent variables.The smaller the evaluation function value is,the more similar the light intensity distribution of the light source model is to the actual light intensity distribution.We use simulated annealing algorithm to find the best equivalent parameters to minimize the evaluation function value.By this method,we establish the light equivalent model of OSRAM two LEDs and get the light intensity distribution of the model. The normalized cross-correlation coefficient NCC between the intensity distribution and the actual LED intensity distribution is above 99%, which confirms the feasibility of this method.
The effects of phosphor layer concentration and thickness on the light intensity distribution and spatial color temperature distribution of LED are analyzed.The parameters of phosphor can be adjusted accurately by using the optimization algorithm of neural network and simulated annealing,so that the spatial color temperature distribution of the output light is close to the predetermined spatial color temperature distribution.Taking Ruifeng photoelectric chip 3535 as an example,the system is on. The equivalent model of phosphor obtained by this method is very similar to the chip's spatial color temperature distribution, and the NCC is as high as 99.1%. Using the same method, a photochromic composite mode light source is established. The light source of Ruifeng 3535 chip is modeled equivalently. The NCC of the equivalent light-color composite model and the actual chip is 97.5% and the space color temperature distribution is 97.5%. It is shown that the light-color composite model can be very similar to the actual LED light-color distribution. The method can simulate white LED light source better.
Based on the equivalent white LED light source, a lens is designed to make the light produced by the white LED light source produce uniform illumination and chromaticity distribution on the illumination target surface. So we can only consider one contour of the free-form surface when we design. The contour curve of the free-form surface adopts cubic spline curve. The initial structure is adjusted manually to make the illumination and spatial color temperature distribution close to the ideal illumination distribution and spatial color temperature distribution. The interactive optimization function was further optimized. The uniformity was 93% and 91% respectively. The uniformity was increased by 17% and 8% respectively compared with the illumination and spatial color temperature distribution produced by the lens before optimization.
【學(xué)位授予單位】:江南大學(xué)
【學(xué)位級別】:碩士
【學(xué)位授予年份】:2014
【分類號(hào)】:TM923.34
【參考文獻(xiàn)】
相關(guān)期刊論文 前1條
1 蔣大鵬,趙成久,侯鳳勤,劉學(xué)彥,范翊,張立功,褚明輝,申德振,范希武;白光發(fā)光二極管的制備技術(shù)及主要特性[J];發(fā)光學(xué)報(bào);2003年04期
本文編號(hào):2216102
本文鏈接:http://sikaile.net/kejilunwen/dianlilw/2216102.html
最近更新
教材專著