Currently, the beginning of transition to emergency states is not reflected in the readings of measurement equipment of control systems of many essential technical facilities. The signaling of the beginning of the latent period of an emergency state is not implemented either. There are no tools for monitoring malfunctions of the railroad bed, or for controlling the beginning of the emergence of various typical malfunctions in the equipment of urban water supply systems, which are operated in continuous rotary motion under high pressure. Control systems of said facilities detect malfunctions only when they already have a clearly pronounced form. Nevertheless, specialists and operating personnel use this information to take action and prevent various accidents. However, in some cases, it is too late and an accident becomes unavoidable. At the same time, the emergence of a malfunction at these facilities is accompanied with the emergence of noise in the vibration signals at the outputs of the vibration sensors in their control systems. This noise contains diagnostic information. In this study, we propose technologies for generating informative attributes that can be used to control the beginning of the latent period of accidents. We also show the possible applications of the proposed technologies in developing intelligent systems for controlling the onset of accidents at facilities under investigation and other strategic construction.