OPTIMIZATION PROBLEM FOR LIGHT QUALITY CONTROL OF POLYCHROMATIC SOLID-STATE LIGHTING DEVICES

Ignas Dapsys1, 2

1 Department of Mathematics and Informatics of Vilnius University, Naugarduko St. 24, LT-03225 Vilnius

2 Baltic Institute of Advanced Technology, Sauletekio Ave. 15, LT-10224 Vilnius

Solid-state lighting (SSL) offers superior energy efficiency and control possibilities than traditional light sources such as incandescent and compact flourescent light bulbs [1]. Therefore, many groups of scientists around the world are interested in research and development of SSL. Among such research is the study of colorimetric properties of artificial light [2], its psychophysical effect to an observer and the control of light quality to maintain human needs [3]. Multiple ways have been devised to numerically define color rendering, such as CRI and CQS [4] for human vision, as well as TLCI [5] - for CCD sensors of digital cameras. Therefore, the color rendering is ambiguous, whereas lighting industry requires adaptive lighting fixtures with optimal rendering quality for not just human vision. Although SSL offers better energy savings, their color rendering and adaptive control of color temperature leave much to be desired. Yet, it may be overcome by the control capabilities offered by polychromatic SSL. By changing the electrical current driving a SSL device, we can modify the light output, i.e. illuminance E, correlated color temperature T. In this work, a mathematical model for a control of polychromatic SSL device, that will be used to maximize color rendering index R while changing the correlated color temperature T and maintaining constant illuminance E is going to be presented.

The modelled polychromatic SSL device consists of n light emitting diodes (LEDs), each of different color, and is driven by a pulse wave modulation (PWM) power source with n independently controllable channels. Each of the channels is described by the control parameter q, that is the pulse width in time scale. The light spectrum of each LED is represented by a Gaussian distribution

$$S(\lambda, q) = P(q) \exp\left\{-\frac{(\lambda - \lambda_{\text{max}})^2 4 \ln 2}{\Delta \lambda^2}\right\}\tag{1}$$

where P(q) is the optical power of an each LED, λmax is the peak wavelength and Δλ is the spectral width at half-maximum. One can see that only the optical power depends on the q. The entire spectrum of this SSL device is given by a sum of the spectra of each LED.

The spectrum can then be used to calculate the illuminance E, color temperature T and color rendering index R. In this case, all depends on n variables: E(q1,q2,..., qn), T(q1,q2...,qn) and R(q1,q2,...,qn). R(q1,q2,...,qn) is the objective function that we seek to maximize by changing the pulse widths q1,q2,..., qn, each for one of the n LEDs, with respect to following conditions: the illuminance and the correlated color temperature must remain constant to the preset values. Therefore, we formulate the following optimization problem

$$\max_{q_1,q_2,\dots,q_n} R(q_1,q_2,\dots,q_n)\tag{2}$$

such that the following constraints are met

$$E(q_1, q_2, \dots, q_n) = E_0, \quad E_0 \in [0, 1],\tag{3}$$$$T(q_1, q_2, \dots, q_n) = T_0, \quad T_{\text{min}} \le T_0 \le T_{\text{max}}.\tag{4}$$

The optimization problem is solved by using the brute force algorithm. The grid with 10 points per control parameter is created. Then values of E, T, and R are calculated with all possible n-tuples of control parameters. That is 10n number of values for each of the function. However only defined number of control parameters can meet the constrains of the optimization problem. The problem is solved with n = 2,3, and 4. The relationship between the color temperature T and color rendering index R is determined. The results of the optimization problem for four LEDs using three polychromatic SSL configurations: RGBA, RGB[WW] and RGB[CW], where R is red, G - green, B - blue, A - amber, WW - warm white and CW - cool white LEDs, are compared.


[1] E. F. Schubert, J. K. Kim, Solid-state light sources getting smart, SCIENCE 308, 12741278 (2005).

[2] A. Zukauskas, R. Vaicekauskas, F. Ivanauskas, R. Gaska, M. S. Shur, Optimization of white polychromatic semiconductor lamps, Appl. Phys. Lett. 80(2), 234 (2002).

[3] M. Aldrich, N. Zhao, and J. Paradiso, Energy efficient control of polychromatic solid-state lighting using a sensor network, Proc. of SPIE 7784, 778408 (2010).

[4] M. Wood, CRI and the Color Quality Scale, Part 2, Protocol 15(2), 1620 (2010).

[5] European Broadcasting Union, Television Lighting Consistency Index, https://tech.ebu.ch/tlci-2012, [accessed 2015-12-16] (2012).