The extensive amount of knowledge possessed by an organization represents one of its most valuable resources. Nonetheless, it is not always easy to pinpoint, arrange and reuse such knowledge in an effective and productive manner. Moreover, in the last two decades companies have been increasingly using serious games and simulations for the most diverse scopes. While a consequent interest has been growing in the field of knowledge management regarding how to codify and reuse the information generated by these new innovative tools, there is still no clear understanding of what are the games' requirements which are responsible for the generation of knowledge. In addition to this, many theories can be identified in the available literature which developed methods and models for knowledge conversion and sharing; however, none of this models has yet been applied to the context of serious games and simulations. As a consequence of this recognized gap in the available literature, as well as of the need for organizations to develop a series of knowledge management systems and practices for the sharing and reusing of knowledge generate through games, this thesis research aims at understanding what are the necessary requirements for games to foster the development, as well as the reuse, of tacit knowledge. Primary scope of this thesis project has thus been to determine what games' elements represent a fundamental requirement for the generation of both explicit knowledge and tacit skills. In addition to this, the research has taken into consideration different groups of actors (game designers, participants/players, project team members, managers), and attempted to verify whether any dissimilarities exist among these groups in the way of perceiving games' requirements, and their impact on knowledge creation and sharing. Finally, in the light of the obtained results, suggestions have been proposed on the most appropriate knowledge management systems and procedure to use in the context of serious games and simulations. The research has been carried out in the form of an embedded, single-case study: the main unit of analysis - and client - was ProRail, the government organization which manages the maintenance and extensions of the national railway network infrastructure, the allocation of rail capacity, and the traffic control. The embedded units of analysis were instead represented by different games employed within the organization (OV-SAAL, Ketensimulatie, ERTMS). Once the games have been identified, other research methods have been applied within the case study framework, in particular: semi-structured interviews and Q-sorting. Concluding this dissertation report, recommendations for ProRail - the customer of this thesis project - have been outlined, by combining the obtained insights with theories from available literature. The scope has been to define suggestions for a knowledge management framework for serious games and simulations. The suggested framework, result of the joint efforts of different TU Delft researchers and ProRail's employees, is presented in the second-last chapter of this dissertation in the form of a set of recommendations for ProRail.