DEVELOPMENT OF AN INTELLIGENT ROBOTIC SYSTEM FOR MULTISENSOR MONITORING AND DIAGNOSTICS OF ENERGY INFRASTRUCTURE

The article examines the scientific and practical foundations associated with the implementation of robotic technologies for predictive maintenance in electric power systems. The increasing complexity of power infrastructure underscores the significance of this study, highlighting the escalating demands for reliability and continuity of electricity supply, as well as the need to mitigate risks to maintenance personnel during inspections of high-voltage, hard-to-access, or potentially hazardous facilities. The discussion emphasizes the application of mobile robotic platforms, unmanned aerial vehicles, robotic manipulators, and autonomous diagnostic systems for the surveillance of power lines, transformer substations, switchgear, solar power plants, and other critical components of electric power systems. The objective of this study is to substantiate the architectural framework and methodological principles of a robotic predictive maintenance system that incorporates multisensor data acquisition, intelligent diagnostics, condition assessment, and decision support for maintenance planning. The proposed methodology employs visual, thermal, acoustic, vibration, electrical, and environmental data gathered by robotic platforms during scheduled or autonomous inspection missions. This article delineates the principal stages of data processing, encompassing sensor synchronization, preprocessing, defect detection, feature extraction, anomaly identification, classification of technical conditions, and prediction of the remaining useful life of power equipment. A structural and functional model of a robotic predictive maintenance system is presented, integrating field-level robotic inspection tools, edge computing modules, communication channels, cloud-based analytical services, and interfaces with SCADA, CMMS, GIS, and digital twin platforms. The application of artificial intelligence techniques, including computer vision, machine learning, deep neural networks, and time-series analysis, is justified for detecting overheating, insulation degradation, mechanical damage, corrosion, loose contacts, conductor defects, oil leakage, and other early-stage failure indicators. This study demonstrates that robotic technologies can substantially enhance the efficiency of predictive maintenance within electric power systems by increasing both the frequency and accuracy of inspections, minimizing human involvement in hazardous operations, facilitating remote diagnostics, and promoting data-driven maintenance planning. The proposed concept provides a methodological foundation for developing intelligent robotic diagnostic systems to improve the reliability, resilience, and operational safety of contemporary electric power infrastructure.