Nowadays the wide integration of automation resources is occurred in many domains of industry. One of the key aspects of this process is control automation for breaking up process on railways. Due to the fact that main element of railway yard is hump the Intellectualization of decision making processes becomes the key task in area of fault prediction during breaking up. The experience of fault effects elimination shows this task decision allows to reduce the financial and time costs for breaking up process. This paper proposes new algorithm for creation of theoretical base of the system for fault prediction. This algorithm is realized by hybrid methods, which combine different temporal models characterizing technological process of sorting of following cuts. Presented experimental arguments show integration of proposed methods is efficient in domain of railway nreaking up process

Keywords: breaking up process, hump yards automation, intellectualization of technological processes, fault detection, hybrid techniques, fuzzy temporal relations, reconstructed phase space