EXPERT SYSTEM FOR ANALYSIS OF ACCIDENTS IN SHIP BOILER BOILERS INSTALLATION

Abstract

The operation of the ship boiler plant is associated with a large number of potential failures, which, in turn, cause accidents and can lead to human casualties and damage to ship equipment, etc. The paper summarizes data on accidents of ship boiler plants ( explosions, fires, power outages or other serious disturbances in the operation of ships) in order to determine the worst case scenarios during their operation. For the quantitative analysis of the risks of accidents in the ship's boiler plant, the approach using a fuzzy expert system based on the use of interval fuzzy sets of the second type in the relevant production rules has been improved. The areas of uncertainty of interval fuzzy sets of the second type are used as membership functions of the
improved fuzzy expert system.

Interval fuzzy sets of the second type, on the one hand, provide formalization of more additional degrees of uncertainty in comparison with classical fuzzy sets of the first type, on the other hand, are realized in development of fuzzy logical systems and have less computational complexity, in comparison with fuzzy sets of another type.

During the study, three types of membership functions were tested to create fuzzy sets and distinguish between these functions (Gaussian membership function (GMF), triangular membership function (TMF) and trapezoidal membership function (ZMF)).

As a basic methodology for accident analysis in the work using the risk analysis method, it considers problems related to system characteristics and failure modes for each component of the system, and ensures the availability of appropriate security measures, which in turn is suitable for identification, identification and verification mechanical and electrical hardware systems for potential failures (Failure Mode and Effects Analysis (FMEA)).