Architectures of engineering support for the product life cycle based on a risk-oriented approach to quality and end-to-end data traceability in the context of Industry 4.0

Abstract

The paper substantiates an architecture of engineering support for the product life cycle in construction and mechanical engineering in the context of Industry 4.0, based on a risk-oriented approach to quality and end-to-end data traceability. The problem of «information gaps» between the stages of design, construction, and operation is formulated, leading to an increase in data entropy and a sharp rise in the probability of critical failures and defect correction costs. The methodological framework includes systems analysis of life cycle processes, object-oriented modeling and graph theory, the analytic hierarchy process with fuzzy logic, Markov reliability models, Monte Carlo simulation, and NLP methods for processing unstructured data; the empirical base comprised 48 facilities, more than 45 TB of information, high-LOD BIM models, and a sample of over 1.5 million structural elements with digital quality passports. Quantitative dependencies have been obtained between the level of end-to-end traceability implementation and the probability of critical risks at different life cycle stages: a more than twofold reduction in integrated risk and maximum effect at the operation stage have been demonstrated (decrease in probability from 0.521 to 0.194 with an efficiency coefficient of 2.685), and the key role of the quality of attribute data on operational requirements has been identified, for which attribute losses reach 14.89% and have the highest correlation with installation errors. It has been shown that the transition from the basic CAD level to the ecosystem level «BIM Level 3 + IoT» reduces the specific cost of error correction by 5.7 times, decreases the payback period of the system to 14.6 months, and increases the integral quality indicator to 0.943, confirming the high investment attractiveness of an integrated digital architecture. Requirements for ontological data models, mechanisms of online verification and validation of information flows, as well as prospects for scaling the proposed solutions to the level of urban digital platforms and smart city systems are discussed.