This paper presents a proposal for structuring events for system models expressed using IOPT nets (Input-Output Place-Transition Petri nets). Currently, a non-autonomous event within an IOPT model is defined based on change of input signals with respect to a specific threshold, when two consecutive execution steps are considered. New types of events are proposed, allowing the definition of an event activated not only by crossing a fixed threshold, but also by considering a change in associated signal values on a specific amount (belonging to an interval of values). The concept is further extended allowing the definition of an event based on signal values presented on previous execution steps. The proposal results on a classification of several types of events, namely threshold events, momentum events, impetus events, as well as delayed events and logical events. Usage of these types of events allows improvements in terms of expressiveness and compactness of the resulted model.