%0 Trabajo de grado (Bachelor Thesis) %A Quintero Sánchez, Laura Camila %D 2019 %G Desconocido (Unknown) %T Numerical predictions of sprinkler activation times and comparison with semi-empirical correlations = Predicciones numéricas de los tiempos de activación de los rociadores y comparación con correlaciones semi-empíricas %U http://babel.banrepcultural.org/cdm/ref/collection/p17054coll23/id/1262 %X This research contrasts two of the most common methods used to estimate sprinkler response, semi-empirical correlations and computational models based on fluid dynamics. The semi-empirical calculations are based on Alpert’s relations and the modelling task is carried out with the software tool Fire Dynamic Simulator, FDS. The case studies implemented in this work were defined considering the limitations of Alpert’s equations. Different enclosure configurations were considered, such as, unconfined spaces (open rooms - steady state and time-dependent fires), confined spaces (rooms with parallel walls creating corridors – steady state fires) and obstructed areas (rooms with beams in the ceiling – steady state fires). The results obtained show that both approaches predict accurately the sprinklers activation time for unconfined configurations, except for the heat release of 500 kW since Alpert's relationships are not applicable for small fires. In general, the tendency observed is that semi-empirically calculated values are more conservative, resulting in the prediction of longer activation times. Therefore, this would result in designs based on the worst-case scenario, showing that Alpert’s correlations can be applied with certain confidence. On the contrary, for confined and obstructed scenarios, the outcomes show that the semi-empirical correlations are not completely reliable for this type of enclosures. In confined configurations, regardless of the magnitude of the heat release rate, the more confined the corridor, the sooner the sprinkler activation occurs. Therefore, in these cases, although Alpert's relationships consider the worst risk scenario by predicting longer activation times, the design of fire protection systems will be based on inappropriate values. In obstructed scenarios, even the smallest obstruction modifies the ceiling jet flow due to the presence of the beams, varying also the smooth ceiling jet attributes characteristic of unconfined scenarios. Therefore in rooms with obstructions, the semi-empirical correlations are not applicable anymore for devices located in a bay other than the primary, since Alpert´s equations predict shorter activation times to those that could be expected in reality, which means that the worst risk scenario would not be considered. This study demonstrates the importance of using computational tools for sprinkler response estimations, being a useful and valid alternative to analyse complex fire scenarios and to improve fire protection systems designs.