Development of An Ai-Driven Predictive Maintenance and Optimization Framework for Industrial Compressor Systems

This study examines failure modes of multi-stage compressors and explores optimization strategies to enhance their performance in power plants. Using Root Cause Analysis, MTBF, and Weibull models, the research identifies pressure-related failure trends, with HP stages prone to material stress and vibration, and IP stages affected by high temperatures and wear. Despite lower pressures, LP stages also experience structural damage. The findings advocate for tailored maintenance approaches, improving system reliability, reducing costs, and minimizing environmental impact. Optimized systems showed significant improvements in MTTF, MTTR, and system availability, reaching 95.149%. Shortened maintenance intervals increased CO₂ emissions, highlighting the environmental trade-offs of over-servicing. Machine learning was employed to optimize maintenance timing, reducing emissions and delays. Overall, the research underscores the importance of reliable, sustainable, and cost-effective compressor operations for industry and environmental sustainability.