Non-Parametric classification of technical condition parameters based on shift and scale tests

While creating contemporary management systems and automating technology processes and objects, there are different methods of signal processing based on computer and information technologies that can be employed. An important research direction when creating the systems is the research direction of the reliability and assessment of technical condition of various objects in the industrial engineering field. For assessing the object technical condition, various technical condition parameters are used. In the face of uncertainty, the parameters are considered as random processes. In practice, the technical condition parameters are values of time series, their quality of analysis and forecasting largely depends on their numeric statistical characteristics and the most important of them is stationarity (non-stationarity). The paper proposes a non-parametric classification algorithm of the different technical condition parameters. The algorithm classifies all technical condition parameters into four classes based on the Kruskal-Wallis shift test and the Fligner-Killeen scale test.