ARTIFICIAL NEURAL NETWORK MODELLING OF COLOUR TEMPERATURE VARIATIONS WITH DIFFERENT TYPES OF ARMATURES AND LIGHT SOURCES

Abstract

The continually growing demand for energy and the related resource problems have increased the importance of detailed consumption habits. A significant percentage approximately 20 % at present of the total electrical energy consumed is used in residential areas, in the work-place, and for street and road lighting. In the past, lighting was mainly used to improve visibility and safety, but efficiency and aesthetics are now playing a big role in the lighting sector, hence different types of lamps are now sold on the market. In this study incandescent, compact fluorescent and LED type lamps, which are widely used among the customers, were selected as different types of light sources. To represent the different armature types transparent glass prisms were prepared with different thickness and shapes. Initially the colour temperature of incandescent, compact fluorescent and LED type lamps was recorded under laboratory conditions. The aim of this study is to model an artificial neural network (ANN) for estimating colour temperature values from obtained input values. When obtained and estimated data are compared, it is observed that the estimation process and method is successful for this type of data.