Information and Analytical Support for Management of Fire Extinguishing at Highly Dangerous and Technically Complex Facilities

Abstract

In this article the authors formulate the task for process management of man-made fire extinguishing at highly dangerous and technically complex facilities. This task consists of fire localization and elimination using minimal assignment in minimum amount of time. For this task the authors developed a neuro-fuzzy model for fire extinguishing process control, the main elements of which are a neuro-fuzzy model for predicting the fire area, a neuro-fuzzy model for selecting the fire rank, a neuro-fuzzy model for evaluating the implementation success of the plan, a neuro-fuzzy model for selecting the optimal action plan, an analytical model for evaluation of resources sufficiency, an analytical model for resources selection, and a model for implementation of neuro-fuzzy models. In comparison with existing models, distinctive features of the developed model are the following: application of combined (bell-shaped with thresholds) membership functions that allow to perform more accurate approximation of input parameters values; implementation of the block to eliminate dynamic errors. This paper assesses model adequacy through verification and validation. The authors developed a system for fire extinguishing process control. This system allows us to raise of firefighters' efficiency due to increase of accuracy of managerial decisions taken by the manager and time reduction needed to formulate a decision.