Glare and visual discomfort are important factors that should be taken into account in illumination design. Conventional freeform optics offer perfect control over the outgoing intensity distribution, but few control over the near-field luminance distribution. Observing the emitted light from a high-brightness LED through a freeform lens, yields a high peak luminance that can result in glare. Diffusors can be used to reduce these peak luminances, but such components do not offer accurate control over the resulting intensity pattern. An alternative approach to reduce the observed peak luminance is by spreading the emitted light over multiple optical channels via freeform lens arrays. Such freeform arrays still allow excellent control over the resulting intensity distribution. In this paper the design of freeform lens arrays is discussed in both 2D and 3D. The possibilities and main limitations of the resulting components are discussed in detail by considering a few examples. In order to understand the design algorithms for multi-channel freeform optics, the paper starts with a brief introduction to the design of single-channel freeform optics for obtaining prescribed illumination patterns.
Given the high source luminance of current white LEDs, spreading this luminance in order to avoid glare, is a crucial aspect of effective LED lighting system design. In previous work, a method has been presented to design a rotational symmetric or extruded lens consisting of different freeform segments. These segments are designed in such a way that they spread the incoming rays into multiple overlapping fans of which the combination forms a desired target intensity pattern. The advantage of using freeform lens segments is that these can be precisely tailored to different target intensity patterns, e.g. intensity patterns with a sharp cut-off. The transformation of the incoming intensity distribution to the target intensity distribution can be seen as a linear transformation. If the source and target distributions are discretized into a finite number of bins, it is possible to represent this transformation as a matrix. With an iterative procedure one can quickly generate such transformations for specific source and target distributions. The spreading of light at each point of the optical surface can then be realized with small freeform lens segments. These individual freeform segments can be implemented in two different ways, convex or concave. In this paper, the case is considered in which convex and concave freeform structures are alternated which allows for the creation of a continuous, smooth, oscillating lens surface. This approach can improve both the performance and manufacturability when compared with a discontinuous surface. A drawback of the approach is the fact that one loses control over the overall form factor of the resulting lens. The derived iterative procedure was used to design a luminance spreading illumination lens with smooth, wavy structures on one side and a flat surface on the other side. The lens was designed to generate a specific wide-beam target intensity distribution when combined with a high-brightness LED.
KEYWORDS: Modeling, Light sources, Waveguides, Reflection, Scattering, Light scattering, Clouds, Monte Carlo methods, Ray tracing, LED backlight, Optimization (mathematics), Testing and analysis, Performance modeling, Light
A key requirement to obtain a uniform luminance for a side-lit LED backlight is the optimised spatial pattern of structures on the light guide that extract the light. The generation of such a scatter pattern is usually performed by applying an iterative approach. In each iteration, the luminance distribution of the backlight with a particular scatter pattern is analysed. This is typically performed with a brute-force ray-tracing algorithm, although this approach results in a time-consuming optimisation process. In this study, the Adding-Doubling method is explored as an alternative way for evaluating the luminance of a backlight. Due to the similarities between light propagating in a backlight with extraction structures and light scattering in a cloud of light scatterers, the Adding-Doubling method which is used to model the latter could also be used to model the light distribution in a backlight. The backlight problem is translated to a form upon which the Adding-Doubling method is directly applicable. The calculated luminance for a simple uniform extraction pattern with the Adding-Doubling method matches the luminance generated by a commercial raytracer very well. Although successful, no clear computational advantage over ray tracers is realised. However, the dynamics of light propagation in a light guide as used the Adding-Doubling method, also allow to enhance the efficiency of brute-force ray-tracing algorithms. The performance of this enhanced ray-tracing approach for the simulation of backlights is also evaluated against a typical brute-force ray-tracing approach.