Methods of traction rolling stock fire safety analysis

Abstract

Aim. Fire safety of a protection asset is the state of a protection asset that is characterized by the capability to prevent the occurrence and development of fire, as well as the effects of hazardous factors of fire on people and property [1]. The traction rolling stock (TRS) is one of the primary protection assets on railway transport. Managing TRS fire safety involves a large volume of information on various TRS types: possible fire-hazardous conditions, fire safety systems, parameters of TRS-related processes. That means that efficient management must be built upon analysis that allows identifying trends and factors of fire hazard development. The analysis should be organized in such a way as to allow its results to be used in evaluation of composite safety indicators [2]. The required applied nature of such analysis is also obvious. Given the above, it should be noted that the applied research indirectly solves the task of using the results of fundamental research to address not only cognitive, but also societal issues [3]. The aim of this article is to structure the most efficient applied and theoretic methods of analysis and to develop a structure for systems analysis of TRS fire safety. Methods. The multitude of factors that affect the condition of TRS can be divided into two groups: qualitative and quantitative. Importantly, it is impossible to completely research the impact of all the elements of a complex technical system that is TRS on fire safety. We have to examine a part of the whole, i.e. a sample, and then use probabilistic and statistical methods to extrapolate the findings of sample examination to the whole [4]. An analysis of a data set requires a correctly defined sample. At this stage, the quality of information is the most important criterion. The list of raw data was defined based on the completeness of the description, reliability of the sources. Then, in a certain sequence, the data was analyzed by means of qualitative and semi-quantitative methods. First, given the impossibility to establish evident connections (destroyed by the hazardous effects of fire) between the condition of units that preceded the fire, the Pareto analysis was used. The research involved root cause analysis (Ishikawa diagram). Subsequently, cluster analysis of fire-hazardous situations was used. The main purpose of cluster analysis consists in establishing generic sequences of events that entail TRS fires. For that purpose, a description of possible fire-hazardous states of traction rolling stock is required, i.e. a multitude of events and states must be described. Dependability analysis can be successfully performed by representing the safety state information in terms of the theory of sets [5]. The sets of hazardous fire-related TRS events are represented in the form of partially ordered sets. Processing of such sets that are non-numeric in their nature cannot be performed by means of statistical procedures based on addition of parametric data. For that reason the research used mathematical tools based on the notion of type of distance. A part of data that have quantitative characteristics was analyzed statistically. Results. The TRS fire safety data analysis methods presented in this article that include methods of numeric and non-numeric data processing allowed developing a formatted list of fire hazard factors that enable the creation of a practical method of TRS fire risk calculation. An algorithm is proposed for application of qualitative and quantitative methods of analysis of data of various numerical natures. An example is given of the algorithm’s application in the analysis of diesel engine fire safety. The proposed method can be used for analyzing anthropogenic safety in terms of listing the factors involved in risk assessment.