Estimation of efficacy factors and luminance smoothness of road lighting systems with artifical neural networks

In order to save energy in the field of lighting, very simple measures are often sufficient. The most energy loss in lighting is due to the unnecessary use of lamps. When the lighting is their choice, people turn on the lights of the environment they enter, whether artificial lighting is necessary or not. However, when they are done with their work, they may neglect to turn off the lamps when they leave the place. This event is usually seen in places open to common use. Therefore, a large part of the electrical energy used in lighting is wasted without being used for any work. In order to minimize the energy expenditure in the lighting business, it is not a solution to reduce the illuminance level. Rather than doing so, it is to prevent as much as possible the approaching losses and unnecessary expenditures. Due to the increase in energy production costs and the increasing need for energy in every field, it has become a necessity to save on energy. Energy saving in lighting, on the other hand, can be achieved by providing an equivalent level of illumination economically without any decrease in lighting quality. The desired lighting is possible by sending the required amount of light in the areas to be illuminated. Thus, since the areas in use are illuminated as desired, loss of energy will be prevented. With the lighting of the desired quality, less electrical energy consumption will be achieved, as well as a good quality lighting will be provided. The efficiency factor (luminous efficiency) is an important criterion for good lighting. By adjusting the efficiency factor of the lamps, the electrical energy consumed by the lamps can be minimized. The higher the efficiency factor value in a lamp, the less electrical energy will be consumed for lighting. Thus, it will contribute to the economy as well as energy savings. In addition, it is an important criterion for good road lighting to provide sufficient glow uniformity for energy saving in road lighting. When the brightness uniformity is good in road lighting, unnecessary energy expenditures will be avoided. In this thesis, the efficiency factors of these lamps were estimated by using the lamp types of lamps of different brands, the voltage values of the lamps, the luminous fluxes of the lamps and the color temperatures of the lamps. According to the results of the estimation made with the created ANN model, the efficiency factor of different types of lamps was estimated with an accuracy rate of over 98%. The lighting systems to be built will provide the same level of illumination economically without reducing the lighting quality, and it will be easier to decide which lamp with the efficiency factor should be selected in order to increase the light efficiency. In addition, the luminance values on the road surface were measured on a determined road. Prediction was made using artificial neural networks. Thus, it has been possible to estimate the luminance (cd/m2) values of certain points on the road in a short time period and with minimum error. Thus, it will be possible to make the road lighting systems, which cannot be done due to the complex traffic density, and to make the measurement results, which take a long time, in the shortest time and more reliably. Since there is no need to measure at every point, it will be easier to determine whether the lamps used in road lighting are in time for maintenance. Thus, the luminance values of the whole road (cd/m2) will be estimated by Artificial Neural Networks by measuring at certain points determined without hindering the traffic flow on the road. For this, firstly, information about lighting concepts and artificial neural networks is given. Then, measurements were made on the determined road surface. The data obtained in line with the measurements were trained with artificial neural networks. Estimates have been made. The application part of the thesis was carried out in three parts. First, the path to be measured was determined and various calculations were made according to its length and width. As a result of the calculations, the points on the road to be measured are marked. At the marked points, the luminance (cd/m2) values of the road were measured from different points and angles by using the 'tripod luminancemeter' device and the values obtained were recorded. In the second stage, some of the data obtained as a result of the measurements were transferred to the tables and trained with artificial neural networks. As a final step, the consistency in estimating the luminance (cd/m2) value of artificial neural networks with the remaining luminance (cd/m2) data was tested. In the last part of the thesis, the compatibility between the predictions obtained at the points determined by the artificial neural network made for the examined path and the actual values was examined. Comments on the implementation of the bill in practice were also made. Looking at the results, it has been observed that the predictions made with the artificial neural network are close to the truth and are applicable. Keywords: Luminous Efficacy, Energy saving, Lighting quality, ANN