Industrial processes emit large amounts of diverse pollutants into the atmosphere, among which NOx takes a significant portion. Selective non-catalytic reduction (SNCR) is a relatively simple method for the NOx reduction in large industrial facilities such as power plants, cement plants and waste incinerator plants. It consists of injecting the urea-water solution in the hot flue gas stream and its reaction with the NOx. During this process flue gas enthalpy is used for the urea-water droplet heating and for the evaporation of water content. After water evaporates, thermolysis of urea occurs, during which ammonia, a known NO reductant, and isocyanic acid are generated. In order to cope with the ever stringent environmental norms, equipment manufacturers need to develop energy efficient products that are at the same time benign to environment. This is becoming increasingly complicated and costly, and one way to reduce production costs together with the maintaining the same competitiveness level is to employ computational fluid dynamics (CFD) as a tool, in a process today commonly known under the term "virtual prototyping". The aim of this paper is to show capabilities of the developed mathematical framework implemented in the commercial CFD code AVL FIRE, to simulate physical processes of all relevant phenomena occurring during the SNCR process. First, mathematical models for description of SNCR process are presented and afterwards, models are used on the 3D geometry of an industrial reactor and a real industrial case to predict SNCR efficiency, temperature and velocity field. Influence of the main operational parameters on NOx reduction efficiency was performed on the same case. Finally, conclusions about validity of current framework are given together with recommendations for further work. (C) 2016 Elsevier Ltd. All rights reserved.