Artificial neural network application to a building hybrid illumination system

The search for new control systems which present advances in energy consumption programs has increased in the last decades, not only due to the widespread use of microprocessors in the different sectors of engineering but also to the advent of the techniques of artificial intelligence. The purpose of this paper is to show the possibility of the use and to analyze automated systems that make use of the artificial lighting in addition to the natural one. An algorithm based on neural computation was developed to analyze the energetic performance of this system, in a multi floor building employed for a commercial office. The hybrid lighting automated systems shall be analyzed taking in mind the influence of natural factors, constructive details of the building and others that define the energetic efficiency of the building. The use of neural computation makes possible a predictive model to operate these systems, increasing the availability and the reliability of the diagnostics, and is a computational tool with a large performance to the solution of problems that are very complex with the traditional mathematic modeling.