Prevention of Local Emergencies in the Mechanical Transport Systems

The paper presents with the problem of detecting emergency situations in a mechanical transport system and taking measures to prevent them. The mechanical transport system is described by a multi-product network model whose flows have variable intensity. The cause of the emergency is the interaction of the flows on the conveyor. The condition for the occurrence of a local accident situation is given by a logical expression linking the intensities of individual product flows. The condition is formulated by experts. Emergency prevention strategies are selected based on the condition of occurrence. The strategy is to choose a backup path for transmitting a flows or reduce its intensity at the input. The ambiguity and uncertainty of the input parameters and the solutions being formed is overcome by the introduction of a special way of representing knowledge - an image of the threat of an accident. The image includes the previously recorded accident precedent and a set of its permissible transformations. The permissible transformations reflect the analytical experience of the expert, obtained as a result of a posteriori analysis of the precedent. The concept of permissible transformations makes it possible to increase the reliability of decisions made. Algorithms for transforming the experience of analyzing and preventing threats are given. The limits of application of the proposed method are discussed.