INTEGRATED EDGE–CLOUD ANALYTICS FOR PREDICTIVE MAINTENANCE AND ENERGY-EFFICIENT SMART MANUFACTURING

Abstract

The rapid development of smart manufacturing has increased the deployment of Industrial Internet of Things devices, intelligent machinery, cyber-physical production systems, digital twins, machine-learning models, and cloud-based industrial analytics platforms. These technologies generate large volumes of heterogeneous operational data that must be processed efficiently to support predictive maintenance, process optimization, fault diagnosis, production intelligence, and energy management. Conventional cloud-centric manufacturing architectures transfer most industrial data to centralized data centers, creating communication latency, network congestion, bandwidth consumption, privacy concerns, and unnecessary energy expenditure. Pure edge-based systems, on the other hand, provide rapid local processing but are constrained by limited computational capacity, storage resources, model complexity, and longterm analytical capabilities. This paper proposes an Integrated Edge–Cloud Analytics framework for Predictive Maintenance and Energy-Efficient Smart Manufacturing that combines real-time edge intelligence with scalable cloud analytics. The proposed methodology organizes industrial data processing through interconnected sensing, edge intelligence, communication, cloud analytics, predictive maintenance, digital twin, energy optimization, and decision-support components. Industrial sensor streams are initially processed near manufacturing equipment using filtering, synchronization, feature extraction, anomaly screening, and lightweight inference. Only relevant events, compressed features, aggregated measurements, and selected historical data are transmitted to cloud services for computationally intensive model training, fleet-level analytics, long-term trend analysis, and optimization. The framework supports adaptive workload placement based on latency sensitivity, computational demand, network conditions, maintenance criticality, and energy considerations. Predictive maintenance services analyze vibration, temperature, acoustic, electrical, operational, and process measurements to identify early degradation patterns and support maintenance planning. Energy-aware analytics continuously evaluate machinery utilization, computational resource consumption, idle behavior, and production conditions to improve overall sustainability. A comparative analytical evaluation indicates that the integrated architecture can reduce cloud data transmission, average maintenance response time, computational energy consumption, unplanned downtime, and carbon-related impact while maintaining high fault-detection performance. The study demonstrates that coordinated edge–cloud intelligence provides a scalable foundation for sustainable smart manufacturing by combining localized responsiveness, centralized learning, predictive asset management, and energy-efficient industrial operations.