Paper detail

Development of a new model to predict indoor daylighting: Integration in CODYRUN software and validation

Many models exist in the scientific literature for determining indoor daylighting values. They are classified in three categories: numerical, simplified and empirical models. Nevertheless, each of these categories of models are not convenient for every application. Indeed, the numerical model requires high calculation time; conditions of use of the simplified models are limited, and experimental models need not only important financial resources but also a perfect control of experimental devices (e.g. scale model), as well as climatic characteristics of the location (e.g. in situ experiment). In this article, a new model based on a combination of multiple simplified models is established. The objective is to improve this category of model. The originality of our paper relies on the coupling of several simplified models of indoor daylighting calculations. The accuracy of the simulation code, introduced into CODYRUN software to simulate correctly indoor illuminance, is then verified. Besides, the software consists of a numerical building simulation code, developed in the Physics and Mathematical Engineering Laboratory for Energy and Environment (P.I.M.E.N.T) at the University of Reunion. Initially dedicated to the thermal, airflow and hydrous phenomena in the buildings, the software has been completed for the calculation of indoor daylighting. New models and algorithms - which rely on a semi-detailed approach - will be presented in this paper. In order to validate the accuracy of the integrated models, many test cases have been considered as analytical, inter-software comparisons and experimental comparisons. In order to prove the accuracy of the new model - which can properly simulate the illuminance - a confrontation between the results obtained from the software (developed in this research paper) and the major made at a given place is described in details. A new statistical indicator to appreciate the margins of errors - named RSD (Reliability of Software Degrees) - is also be defined.